Strengths and weaknesses of the most machine learning methods discussed here appearing in radiation oncology studies.
\r\n\t[2] J. V. Moloney, A. C. Newell. Nonlinear Optics. Westview Press, Oxford, 2004.
\r\n\t[3] M. Kauranen, A. V. Zayats. Nonlinear Plasmonics. Nature Photonics, vol. 6, 2012, pp. 737-748.
\r\n\t[4] P. Dombi, Z. Pápa, J. Vogelsang et al. Strong-field nano-optics. Reviews of Modern Physics, vol. 92, 2020, pp. 025003-1 – 025003-66.
\r\n\t[5] N. C. Panoiu, W. E. I. Sha, D.Y. Lei, G.-C. Li. Nonlinear optics in plasmonic nanostructures. Journal of Optics, 20, 2018, pp. 1-36.
\r\n\t[6] A. Krasnok, A. Alu. Active nanophotonics. Proceedings of IEEE, vol. 108, 2020, pp. 628-654.
\r\n\t[7] M. Lapine, I.V. Shadrivov, Yu. S. Kivshar. Colloquium: Nonlinear metamaterials. Reviews of Modern Physics, vol. 86, 2014, pp. 1093-1123.
\r\n\t[8] Iam Choon Khoo. Nonlinear optics, active plasmonics and metamaterials with liquid crystals. Progress in Quantum Electronics, vol. 38, 2014, pp. 77- 117.
\r\n\t
Early and accurate diagnosis of patients with cerebral demyelinating or infection diseases, space occupying mass lesions and neurological deficits, is essential for optimum treatment decision concerning the administration of specific medication or chemotherapeutic agents, radiation therapy and/or surgical resection.
Currently, conventional MR imaging (MRI) is considered to be an established and useful tool in brain disease detection and it is widely chosen as the initial examination step in patients suspected of brain lesions as it is effective in simultaneously characterizing the soft tissue, cerebrospinal fluid (CSF) spaces, and blood vessels. It is a flexible imaging modality for which contrast can be extensively manipulated without patient burdening by ionizing radiation. Nevertheless, the accurate characterization of brain lesions with MR imaging remains problematic in several cases as the sensitivity and specificity with which this modality defines several brain lesions remains limited [1].
To overcome the aforementioned limitation, the development of new imaging techniques is required, in order to highlight functional or metabolic properties of brain tissue. Proton Magnetic resonance spectroscopy (1H-MRS) is one such technique which provides a non-invasive method for characterizing the cellular biochemistry which underlies brain pathologies, as well as for monitoring the biochemical changes after treatment in vivo. It is considered as a bridge between metabolism and the anatomic and physiological studies available from MRI [2].
Until now, 1H-MRS has been used as both a research and a clinical tool for detecting abnormalities -visible or not yet visible- on conventional MRI. Suggestively, Moller-Hartman et al. reported that when only the MR images used for radiological diagnosis of focal intracranial mass lesions, their type and grade were correctly identified in 55% of the cases, however, the addition of MR spectroscopic information significantly raised the proportion of correctly diagnosed cases to 71% [3].
1H-MRS has been always challenging in terms of its technical requisites (field strength, gradients, coils and software), as well as the accurate metabolic interpretation with regards to pathologic processes. However, the clinical applications of 1H-MRS are continuously increasing as the clinical hardware have become more robust and user-friendly along with improved data analysis, spectra post-processing techniques and metabolite interpretation confidence.
The purpose of this chapter is to provide a thorough review concerning the current status of 1H-MRS in terms of its clinical usefulness as well as its technical prerequisites.
In order to introduce the basic concepts and terminology of 1H -MRS, the basic principles of MRS are briefly described below.
Proton is a charged particle with spin, and exhibits the electromagnetic properties of a dipole magnet. When protons are placed in an external magnetic field B0, they align themselves along the direction of the field (either parallel or anti-parallel) and demonstrate a circular oscillation.The frequency of this circular motion (called Larmor frequency) is dependent on the strength of the local magnetic field and the molecular structures at which protons belong. This can be expressed by the Larmor equation:
whereω0 is the Larmor frequency, γ is the gyromagnetic ratio specific for the nuclei, and B0 is the strength of the external magnetic field.
When electromagnetic energy (in the form of a RF pulse) is supplied at this frequency, the molecules absorb this energy and change their alignment. When the RF pulse is switched off, the molecules realign themselves to the magnetic field by releasing their absorbed energy. This released energy is the basis of the MR signal [4].
1H-MRS uses the same hardware as conventional MRI, however, their main difference is that the frequency of the MR signal is used to encode different types of information. MRI generates structural images, whereas 1H-MRS provides chemical information about the tissue under study.
Although recent studies have shown promise for the use of 1H-MRS to investigate malignant processes to prostate [5], breast [6], skeletal muscles [7], cervical and ovarian cancer [8], the overwhelming number of applications have been demonstrated in the brain, due to the absence of free lipid signals in normal cerebrum, relative ease of shimming, and lack of inherent motion artifacts.
The output of 1H-MRS is a spectrum which is described by two axes as it is illustrated in figure 1. The vertical axis (y) represents the signal intensity or relative concentration for the various cerebral metabolites and the horizontal axis (x) serves to describe the frequency chemical shift in parts per million (ppm). The nature of the chemical shift effect is to produce a change in the resonant frequency for nuclei of the same type attached to different chemical species. It is due to variations in surrounding electron clouds of neighboring atoms, which shield nuclei from the main magnetic field (B0). The resulting frequency difference can be used to identify the presence of important chemical compounds. Within the spectrum, metabolites are characterized by one or more peaks with a certain resonance frequency, line width (full width at half maximum of the peak’s height, FWHM), line shape (e.g., lorentzian or Gaussian), phase, and peak area according to the number of protons that contribute to the observed signal. By monitoring those peak factors, 1H-MRS can provide a qualitative and/or a quantitative analysis of a number of metabolites within the brain if a reference of known metabolite concentration is used at a particular field strength [9].
Proton MR spectrum from Parietal White Metter measured at 3T in the normal human brain of a 19-year-old volunteer.
Accurate classification of cerebral lesions by in-vivo 1H-MRS requires determination of the relationship between metabolic profile and pathologic processes.
The assignment and clinical significance of the basic resonances in a spectrum as well as the less commonly detected compounds are discussed below:
N-Acetyl Aspartate (NAA)in 1H-MR spectra of normal cerebral tissue, is the most prominent resonance which originates from the methyl group of NAA at 2.01ppm with a contribution from neurotransmitter N-aspartyl-glutamate (NAAG) (figure 1). NAA is exclusively localized in central and peripheral nervous system and it is synthesized in brain mitochondria. Its concentration subtly varies in different parts of the brain [10] and undergoes large developmental changes, increasing from 4.82mM at birth to 8.89mM in adulthood. Although NAA is considered as a neuronal marker and equate with neuronal density and viability, its exact function remains largely unknown.
The utility of NAA, as an axonal marker is supported by the loss of NAA in many white matter diseases, including leukodystrophies [11], multiple sclerosis (MS) [12] and hypoxic encephalopathy [13], chronic stages of stoke [14] and tumors [1, 2, 9]. However, there are cases when the abnormal levels of NAA do not reflect changes in neuronal density, but rather a perturbation of the synthetic and degradation pathways of NAA metabolism. For instance, in Canavan’s disease high levels of intracellular NAA [15] are due to aspartoacylase (ASPA) deficiency, which is the enzyme that degrades NAA to acetate and aspartate.
Further examples that show the lack of direct relationship of NAA to neuronal integrity include various pathologies such as temporal lobe epilepsy (TLE) [16] or amyotrophic lateral sclerosis (ALS) [17], which exhibit spontaneous or treatment reversals of NAA to normal levels.
Choline-containing compoundscomprise signals from free choline (Cho), phosphocholine (PC) and glycerophosphocholine (GPC), with a resonant peak located at 3.22 ppm. Since the resonance contains contributions from several methyl proton choline-containing compounds, it is often referred as “total Choline” (tCho). tCho is involved in pathways of phospholipid synthesis and degradation thus reflecting a metabolic index of membrane density and integrity as well as membrane turnover [1, 2, 9].
Consistent changes of tCho signal have been observed in a large number of cerebral diseases. Processes that lead to elevation of tCho include accelerated membrane synthesis of rapidly dividing cancer cells in brain tumors [1, 2, 9], cerebral infractions, infectious diseases [18], and inflammatory-demyelinating diseases [19].
Unlike to NAA, which is distributed almost homogeneously throughout the healthy brain, tCho exhibits a marked regional variability with higher concentrations observed in the pons and lower levels in the vermis and dentate [20]. Therefore, detailed knowledge about regional variations of tCho is necessary for an accurate interpretation of the metabolite’s levels, especially in diseases such as epilepsy and psychiatric disorders where tCho is subtly different to normal levels.
Creatine (Cr) and Phosphocreatine (PCr) together they are often referred as total creatine (tCr) because they cannot be distinguished with standard clinical MR unit (up to 7T) and their sum is thus mentioned. Cr and PCr arise from the methyl and methylene protons of Cr and phosphorylated Cr. Within the 1H-MR spectrum, tCr is located at 3.03 ppm and 3.93 ppm resonant frequencies.
In the brain tCr is present in both neuronal and glial cells and is involved in energy metabolism serving as an energy buffer via the creatine kinase reaction retaining constant ATP levels and as an energy shuttle, diffusing from the energy producing (i.e. mitochondria) to energy utilizing sites (i.e. nerve terminals in brain) [21]. As tCr is not naturally produced in the brain, its concentration is assumed to be stable with no changes reported with age or a variety of diseases and is used for calculating metabolite ratios (NAA/Cr, tCho/Cr etc) [21]. Nevertheless, the use of tCr as an internal concentration reference should be used with caution as decreased tCr levels have been observed in the chronic phases of many pathologies including tumors [22], stroke [23] and gliosis [24].
myo-inositol (mI) is acyclic sugar alcohol that gives rise to four groups of resonances with the larger and most important signal occurring at 3.56 ppm. It is observable on short time echo (TE) spectra as it exhibits short T2 relaxation times and is susceptible to dephasing effects due to J-coupling.The exact function of mI is uncertain, however it has been proposed as a glial marker and an increase of mI levels is believed to represent glial proliferation or an increase in glial cell size, both of which may occur in inflammation [2]. Additionally, this metabolite is involved in the activation of protein C kinase which leads to production of proteolytic enzymes found in malignant and aggressive cerebral tumors, serving as a possible index for glioma grading [25]. mI has also been labeled as a breakdown product of myelin. Thus, altered levels of mI have been also encountered in patients with degenerative and demyelinating diseases [12, 15].
Lactate and Lipids, in the normal brain should be maintained below or at the limit of detectability within the 1H-MR spectrum, overlapping with macromolecule (MM) resonances at 1.33ppm (doublet) and 0.9-1.3 ppm respectively. Any detectable increase in lactate and lipids can therefore be considered abnormal. Lactate is present in both intracellular and extracellular spaces and provides an index of metabolic rate and clearance [22]. As an end-product of anaerobic glycolysis, increased lactate levels have been observed in a wide variety of conditions in which oxygen supply is restricted such as in both acute and chronic ischemia [14], metabolic disorders [2], and tumors [1, 2, 9, 22]. Lactate also accumulates in tissues that have poor washout like cysts [26] and normal pressure hydrocephalus [27]. However, in CSF, lactate may be detectable at low levels in normal subjects with prominent ventricles [4].
The spectral region between 0.9ppm and 1.3ppm as referred above; represents the methylene (1.3ppm) and the methyl (0.9ppm) groups of fatty acids. It is during membrane breakdown when fractured proteins and lipid layers become visible. Regardless of the exact molecular source, an elevation of lipid resonances indicates cerebral tissue destruction such as infarction [14], acute inflammation [28] and necrosis [18]. In addition, lipid signals have been observed in patients with several metabolic disorders such as Zellweger syndrome and Refsum’s disease [29].
Glutamate (Glu) and Glutamine (Gln) together they form a complex of peaks (Glx complex) between 2.15 ppm and 2.45 ppm, as their similar chemical structures, renders their distinction difficult within a proton spectra at 1.5T. However, at 3T and above Glu and Gln start to become resolved [30] and at magnetic fields of 7T and higher, the Glu and Gln resonances are visually separated leading to big quantification accuracy [21]. Glu is the major excitatory neurotransmitter in mammalian brain and the direct precursor for the major inhibitory neurotransmitter, γ-aminobutyric acid (GABA). The amino acid Gln, is an important component of intermediary metabolism, is primarily located in astroglia and it is synthesized from Glu [21].
The Glx complex plays a role in detoxification and regulation of neurotransmitters. Increased levels of Glx complex are markers of epileptogenic processes [31] and low levels of Glx have been observed in Alzheimer Dementia and patients with chronic Schizophrenia [32]. Glx complex increment, has been also observed in the peritumoral brain edema correlated with neuronal loss and demyelination [33]. As reported by Malhorta et al., Glx might be used as an in vivo index of inflammation since they observed elevated Glx levels in acute MS plaques but not in chronic ones [34].
Alanine (Ala) is an amino acid present in the normal brain, resonating at 1.47 ppm.It is frequently considered as a specific metabolic charecteristic of meningiomas, however, its identification rate varies from 32% to 100% [3, 22]. It can be also presented in neurocytomas [35], gliomas and PNETs [36]. In vivo 1H-MRS at 1.5T often cannot provide a distinction between Ala and Lac peaks as they resonate in neighboring frequencies. When both metabolites are present they produce a triplet peak located between 1.3 ppm and 1.5 ppm [37] observed at 3T and higher.
Glycine (Gly) is the simplest amino acid and possible antioxidant, distributing mainly in astrocytes and glycinergic neurons, where it is regulated due to its neuroactive properties as an inhibitory neurotrasmitter [28]. It resonates at 3.55 ppm and it overlaps with mI rendering the observation of Gly impossible in a non-processed spectrum. In cases of mI absence, the even low Gly levels can be quantified [38].
High levels of Gly have been observed in glioblastomas, medulloblastomas, ependymomas and neurocytomas [28]. It has also been reported that this metabolite may provide a noticeable metabolic index for the differentiation of glioblastomas from lower grade astrocytomas, primary gliomas from recurrence [38] and glial tumors from metastatic brain tumors [36].
Taurine (Tau)gives two triplets at 3.25 ppm and 3.42 ppm, which can be observed at higher magnetic fields [21] as they significantly overlap with Cho and mI. Tau is an inhibitory neurotransmitter that activates GABA-a receptors or strychnine-sensitive glycine receptors and it has also been proposed as an osmoregulator and a modulator of neurotransmitter action [21]. High levels of Tau have been observed in medulloblastoma, pituitary adenoma and metastatic renal cell carcinoma [39]. Shirayama et al have been also reported increased levels of Tau in the medial prefrontal cortex in schizophrenic patients [40].
Glutathione (GSH)is the major protective molecule of living cells assigned to 2.9 ppm. It serves as an antioxidant and detoxifier thus having an important role against oxidative stress [41]. Glutathione also plays a role in apoptosis and amino acid transport [42].
Altered levels of this metabolite have been reported in acute ischemic stroke patients as ischemia is associated with significant oxidative stress [41], in Parkinson’s disease and other neurodegenerative diseases affecting the basal ganglia [21]. GSH has been also found to be significantly elevated in meningiomas when compared to other tumors [42], showing as well an inverse relationship with glioma malignancy.
Several other amino Acids such asSuccinate at 2.4 ppm, Acetate at 1.92 ppm, Valine and Leucine at 0.9 ppm together with Alanine and Lactate,are the major spectral findings of bacterial and parasitic diseases. Acetate and Succinate are presumably originating from enhanced glycolysis of the bacterial organism [9]. The amino acids Valine and Leukine are known to be the end-products of proteolysis by enzymes released in pus [9]. Specifically, Leucine and Valine peaks have been detected in cystercercosis lesions, however they have not been reported in proton MR spectra of brain tumors [9].
In order to precisely identify the metabolite peaks within a spectrum, several technical considerations should be taken into account concerning the applied magnetic field, the shimming procedures as well as the adequate voxel positioning and the available 1H-MRS techniques, which all highly affect the quality of the yielded spectrum before any post-processing intervention.
In 1H-MRS clinical applications, it is not the signals of water and fat that are of interest, but rather the smaller signals of metabolites, thus a magnetic field of sufficient strength is required. Therefore, most clinical 1H-MRS measurements are performed using MR systems with field strengths of 1.5T and higher. Although more powerful 4-, 6-, 7-, and even 8T MR body scanners are currently in use, the most common high field systems operate at 3T.The main advantage of increasing magnetic field strength is the subsequent increase of the signal-to-noise ratio (SNR). Theoretically, SNR increases proportionally to field strength, however, when put into clinical practice, the study of Barker et al [43], demonstrated a 28% increase in SNR at 3T compared to that of 1.5T at short TEs, appreciably less than the theoretical 100% improvement. Another advantage of magnetic field increment, is the proportional increase of the Chemical Shift, from 220 Hz at 1.5T to 440 Hz at 3T. This is reflected by more effective water suppression and improved baseline separation of J-coupled metabolites such as glutamate, glutamine and GABA, without the need of sophisticated spectral editing techniques [44]. The improvement in spectral resolution is further evident at 7T where weakly represented neurochemicals with important clinical impact, such as scyllo-Ins, aspartate, taurine and NAAG, can be clearly visible [44].
On the other hand, the aforementioned advantages may be hampered by intrinsic field-dependent technical difficulties that should be considered. When the frequency shift between two adjacent nuclei is large enough, a measurable alteration of MR signal, used to encode the x- and y-axis spatial coordinates, will occur producing a spatial misregistration. This means that the volume of MRS information may not be the same as that displayed on the localizer MR image [45]. J-modulation anomalies represent another difficulty encountered at high magnetic fields. The large separation of coupled resonances such as Lactate can result in incomplete inversion of the coupled spin over a large portion of the selected volume, resulting in anomalous intensity losses at long echo times. Strategies to quantify the lactate signal loss have been previously discussed by Lange et al. [46]. Magnetic susceptibility from paramagnetic substances and blood products, are sensibly increased with increasing magnetic field strength. Consequently, magnetic field inhomogeneity and susceptibility artifacts makes more difficult to obtain good-quality spectra, especially from largely heterogeneous lesions [45]. Improved local shimming methods can alleviate the problem.
Shimming refers to the process of adjusting field gradients, either manually or automatically, in order to optimize the magnetic field homogeneity over the volume under study. Magnetic field inhomogeneities result primarily from susceptibility differences between different tissues and between tissue and air cavities, which are scaled non-linearly in ultra-high magnetic fields [47]. Thus, voxels that are placed in inhomogeneous regions of the brain, such as the temporal poles, are difficult to shim due to their close proximity to the sinuses.
Field homogeneity is specified by measuring the full width at half maximum (FWHM) of the water resonance, which determines the spectral resolution. Special emphasis, especially when field is increased, must be placed on shimming, as it increases both sensitivity and spectral resolution. This is why most devices come equipped with second or third order shimming by monitoring either the time domain or frequency domain of the 1H-MRS signal [48]. Some times 4-order shimming might be necessary [49], especially in cases when field homogeneity should be reached in large volumes of interest during magnetic resonance spectroscopic imaging (MRSI).
Effective shimming requires methods for mapping field’s strength variations over the area under study. Methods that have been developed for field mapping can be grouped in two categories: those which are based on 3D field mapping [49] and those which map the magnetic field along projections [50]. In both shimming methods, information about the magnetic field variation is calculated from phase differences acquired during the evolution of the magnetization in a non-homogeneous field.
For a meaningful in vivo 1H-MRS, it is important to locate the voxel in the appropriate region for a reliable metabolic characterization of a lesion [48].
First and foremost, cautious spatial localization is used to remove unwanted signals from outside the ROI, like extracranial lipids and to avoid “partial volume effects”, thereby providing a more genuine tissue characterization. Additional benefits from careful spatial voxel localization, originate from the fact that variations in the main magnetic field and magnetic field gradients, are greatly reduced, thereby providing narrower spectral lines and more uniform proton excitation.
Several lesions and stroke infarcts do not always place themselves in positions that are easy to shim such as temporal lobes, the base of the brain and the cortex near the scull. Small voxels is those regions are easier to shim, but the signal also depends on volume so a voxel with 1-cm sides is often considered the practical minimum size to achieve a reasonable SNR [51].
Spectra can be acquired either with a single voxel (SV) technique (single voxel spectroscopy, SVS) or multiple voxels technique, known as either magnetic resonance spectroscopic imaging (MRSI) or chemical shift imaging (CSI) in two or three dimensions. SVS is based on the stimulated echo acquisition mode (STEAM) [52] or the point resolved spectroscopy (PRESS) [53] pulse sequences while MRSI uses a variety of pulse sequences (Spin Echo, PRESS etc.) [54].
SVS acquires a spectrum from a small volume of tissue located at the intersection of three mutual orthogonal slice-selective pulses as depicted in figure 2. The pulse sequence is designed to collect only the echo signal from the point where all three slices intersect [53].
Schematic representation of the three orthogonal SV slice selective pulses (left) resulting in the signal collection only from the rectangular region of interest.
The advantages of this approach are that:
the volume is typically well-defined with minimal contamination (e.g. extracranial lipids),
the magnetic field homogeneity across the volume can be readily optimized, leading to
improved water suppression and spectral resolution.
The main disadvantage of SVS is that it does not address spatial heterogeneity of spectral patterns and in the context of brain tumors, these factors are particularly important for treatment planning such as radiation or surgical resection.
Lesion’s heterogeneity is better assessed by MRSI. MRSI techniques have been extended to two dimensions (2D) by using phase-encoding gradients in two directions, or, subsequently, three-dimensional (3D) encoding [55]. Thus, the detection of localized 1H-MR spectra from a multidimensional array of locations is allowed (Figure 3). While technically more challenging -due to (1) significant magnetic field inhomogeneity across the entire area of interest, (2) spectral degradation due to intervoxel contamination the so called “voxel bleed”, (3) long data acquisition times and (4) post-processing of large multidimensional datasets- MRSI can detect metabolic profiles from multiple spatial positions, thereby offering an unbiased characterization of the entire object under investigation.
An example of 2D-MRSI of a 50-year old female with a glioblastoma. Simultaneously acquired spectra from multiple regions located at the same plane of the lesion (left). Data are also presented as a metabolic map of Choline/Creatine (right).
Water and peri-cranial lipid suppression techniques are of paramount importance in 1H-MRS procedure in order to observe the much less concentrated metabolite signals. The metabolites of interest are usually about a factor of 8,000 less in concentration than water. Therefore, the water suppression efficiency should be robust and should not vary spatially across the field of view (FOV).
The existing water suppression techniques can be divided into three major groups, namely: (1) methods that employ frequency-selective excitation and/or refocusing pulses; or (2) utilize differences in relaxation parameters; and (3) other methods, including software-based water suppression. The most common method of the first group utilizes multiple (typically 3) frequency-selective, 90° pulses (chemical shift selective water suppression (CHESS) pulses [56], prior to localization pulse sequence. Additionally suppression can be achieved by selectively diphase water, while metabolites of interest are rephased using refocusing pulses during the spin echo period [57]. As water and metabolites T1s are sufficiently different, it is possible to suppress the water signal and observe the metabolites in the close proximity to the water resonance [58]. The third method involves the acquisition of two separated scans in which the metabolite resonances are inverted. The large (unsuppressed) water resonance, as well as the water-related sidebands, is not inverted in either scan. The difference between the two scans therefore results in a water-subtracted (suppressed) metabolite spectrum without any interfering water-related sidebands [21].
Lipid suppression can be performed by avoid the excitement of the lipid signal using STEAM or PRESS localization to select a relatively large rectangular volume inside the brain. Since the extracranial lipids are not excited they do not contribute to the detected signal. Opposite to the strategy employed by volume pre-localization, outer volume suppression pulses (OVS) are applied to presaturate the lipid signal [54]. As illustrated in figure 4, rather than avoiding the spatial selection of lipids, OVS excites narrow slices centered the brain’s lipid-rich regions. Additionally, the difference in T1s of lipids (250-350 msec) and metabolites (1000-2000msec) allows the application of an inversion pulse (inversion time ~ 200 msec), which will selectively null the lipid signal [59]. By choosing the inversion delay such that the longitudinal lipid magnetization is zero, the lipids are effectively not excited.
The location and orientation of OVS pulses have been prescribed in order to saturate as much peri-cranial lipid as possible while the signal within the voxel remains unperturbed.
In MR spectroscopy, post-processing is considered any signal manipulation performed in order to improve the visual appearance of the MR spectrum or the accuracy during metabolite estimation. Therefore, for a reliable analysis of in vivo 1H-MR spectra, an understanding of the principles of post-processing techniques is necessary.
Signal post-processing can be performed either on time domain or after Fourier transformation on frequency domain [60]. Eddy current correction, removal of unwanted spectral components, signal filtering, zero filling, phase correction and baseline correction, consist the most common post-processing techniques for effective signal improvement, and they will be briefly discussed below:
During signal localization RF pulses are applied together with magnetic field gradients. The switching pattern of the gradients applied, can cause eddy current (EC) artifacts that are time and space dependent, causing time dependent phase shifts in the FID and distorted metabolite lineshapes within the spectrum preventing accurate quantification. In a spectrum EC artifacts can be removed by acquiring an additional FID without water suppression. The phase of the water FIDis determined in each time point and it is subtracted from thephase of the corrupted FID [24]. The EC artifact correction comprises the first step of the post-processing procedure.
The removal of unwanted signals from the FID which may disturb signals from the resonances of interest is the next step of signal post processing. A typical example of such an unwanted signal in 1H-MRS is that of water. Water suppression during measurement is never perfect and a residual water signal remains in the spectrum which often has a complicated lineshape [24]. Residual water elimination from the FID can be achieved, either by approximating the water signal and subtract it from the FID, or by eliminate it using special filters [61], or by applying baseline correction for the removal of the broad water peak from the spectrum [62].
The existence of a distorted spectral baseline hampers quantitative analysis as the estimation of metabolite peak areas is not reliable. The main sources of the baseline signal are fast decaying components with very short T2* values such as macromolecules, hardware imperfections, signal from the sample and as mentioned above, inefficient water suppression. Thus, for robust data acquisition and quantification methods, baseline correction is of paramount importance. Delayed acquisition (e.g. TE >80 ms) removes the macromolecules due to their shorter T2 relaxation times (∼30 msec), at the expense of loss of information of many scalar-coupled resonances [21] which have been suggested valuable for tumor and stroke characterization [4, 21, 22, 24, 25, 33].
Special functions, called filters, can be subsequently applied at the signal in the time domain. The goal is to enhance or suppress different parts of the FID leading to improved signal quality. The three most commonly used filtering approaches are: sensitivity enhancement, to reduce the noise from the FID; resolution enhancement, to achieve narrower metabolite linewidths; and apodization for signal’s ripple (due to signal truncation) reduction [62].
The FID of a spectrum, when acquired, is sampled by the analog-to-digital converter over N points in accordance to the Nyquist sampling frequency. Therefore, if the number of points is not sufficient, the reliable representation of the signal fails. Instead of increasing the acquisition time with the inevitable noise increment, the acquired FID can artificially be extended by adding a string of points with zero amplitude to the FID prior to Fourier Transformation, a process known as zero filling. Zero filling does not increase the information content of the data but it can greatly improve the digital resolution of the spectrum and helps to improve the spectral appearance [21], rendering it an important post-processing step.
After Fourier transformation, the spectrum will be phase corrected. When the zero-phased FID signal shifts to the frequency domain, yields a complex spectrum with absorption (real) and dispersion (imaginary) Lorentz peaks. However, when the initial phase is non-zero, it is not attainable to restore pure absorption or dispersion line shapes and phase correction must be applied [4, 21, 62]. A zero-order phase correction compensatesfor any mismatch between the quadraturereceive channels and the excitation channels to produce the pure absorption spectrum, whereas, a first-order phase correction compensates for the nuclei dephase due to the delay between excitation and the detection of FID [62].
The effective differential diagnosis of brain lesions using 1H-MRS depends on the ability of the experienced neuroscientist to interpret and evaluate the metabolic criteria and data underlying each disease. However, similarities in the chemical composition among diseases and/or atypical metabolic characteristics, often burden the diagnosis. Thus, a clinical guide to the main MR spectroscopic findings of cerebral disorders is necessary.
This section focuses on the metabolic patterns of a variety of intra-cranial diseases.
Multiple Sclerosis (MS) is an auto-immune inflammatory disease of the central nervous system (CNS) in which the myelin sheaths around the axons are damaged leading to demyelination, neuronal affection, inflammation, gliosis and axonal degeneration [14]. 1H-MRS is particularly informative in MS, by providing evidence of the two primary pathologic processes of the disease: active inflammatory demyelination and neuronal injury in both lesional and non-lesional brain tissue [63, 64].
Acute demyelinating lesions reveal increased Cho and Lac resonance intensities due to the release of membrane phospholipids during active myelin breakdown and the impaired metabolism of the inflammatory cells, respectively [63]. Short TE spectra also provide evidence of increased lipids, mI [63, 64] and glutamate levels [34]. Increased glutamate levels in acute MS lesions address a link between the direct axonal injury and glutamate excitotoxicity [65], whereas mI is suggestive of glial proliferation and astrogliosis [63]. The aforementioned changes are accompanied by a substantial decrease in NAA due to axonal injury reflecting metabolic or structural changes [64, 65]. It is important tonote that the spectroscopic changes seen in acuteMS plaques are often very similar to the spectraobserved in brain tumors (high Cho, lowNAA, increased Lac, etc.), and therefore this shouldbe kept in mind when evaluating spectra frompatients with undiagnosed brain lesions.
After the acute phase transition, Lac, Cho and lipids seem to return to normal levels, whereas NAA may remain decreased or show partial recovery, lasting for several months [64]. The recovery of NAA can be attributed to resolution of edema, diameter increment of the previously shrinked axons, as a result of the re-myelination and reversible metabolic changes in neurons [64, 65]. There are reports of elevated Cho resonance in chronic MS plaque, probably reflecting the associated gliotic process [66]. Cr seems to be a variable metabolite both in chronic and acute, but is also described to be slowly increasing over time, indicative of gliotic reaction or attempts of incomplete re-myelination of the chronic diseased tissue phases [14].
Metabolic abnormalities in MS patients not only concern the lesions, but are found throughout the normal appearing white matter (NAWM) with notably reduced NAA, which is thought to indicate diffuse axonal dysfunction or loss. It must also be stressed out that the presence of intense gliosis may also cause increased levels of mI [67] and Cr [68]. Increased glutamate, lipids and Cho can be also found in regions of the NAWM, which later are going to develop T2-hyperintense focal lesions [64].
Brain abscesses are focal, intracerebral infections that begin with a localized region of cerebritis, evolving into a discrete collection of pus surrounded by a well-vascularized capsule. The causative organisms involved in brain abscesses are quite variable, and may consist of mixed cultures: aerobes, anaerobes, facultative anaerobes, and facultative anaerobes in combination with aerobes/anaerobes.
MRS has been proven beneficial in differentiating between brain abscesses and other cystic lesions [69], which can be used to implement the appropriate antimicrobial therapy. Brain abscesses reveal specific metabolic substances, such as succinate, acetate, alanine, valine, pyrouvate, leukine, lipids and lactate [69, 70],which are all present in untreated bacterial abscesses or soon after the initiation of treatment [70]. Increases in lactate, acetate, and succinate presumably may originate from the enhanced glycolysis and fermentation of the infecting microorganisms. Amino acids such as valine and leucine are known to be the end products of proteolysis by enzymes released by neutrophils in pus [14]. However, cerebral abscesses contain no neurons [71], therefore no peaks of NAA and Cr/PCr should be detected. The detection of any NAA and/or Cr/PCr is indicative of either signal contamination or erroneous interpretation of acetate peak as NAA [72]. Similarly no tCho peak is present in an abscesses spectrum because there are no membranous structures in its necrotic core [73]. On the other hand, abscesses of tuberculous origin are characterized by the predominant presence of lipids, moderate increase of tChoresonance and no evidence of cytosolic amino acids [4].
Differential diagnosis of brain abscess versus brain tumor is sometimes difficult on the basis of imaging findings and clinical judgment, especially in the case of a brain tumor with a mainly cystic or necrotic component. However, because the vast majority of the aforementioned amino acids have not been detected in brain neoplasms, their presence strongly differentiates abscesses from highly aggressive tumors [71].
Most studies of 1H-MRS of the human brain have focused on the signals from NAA and lactate, as potential markers of brain ischemia, respectively, although there are also often changes in the other metabolite signals, such as Cho, Cr, glutamate (Glu) and glutathione (GSH)[74]. The time course of these metabolite changes through time is an important factor for the diagnosis and prognosis of a brain infarct.
In acute stroke the infarct core rapidly shows signs of cell death and a spectrum from this area has the characteristic lactate peak, often with a broad lipid peak too. Lactate could also arise from a shift toward anaerobic glycolysis in potentially viable cells that continue to metabolize glucose under locally hypoxic conditions [75]. Lactate may also be present in smaller concentrations in the ischemic penumbra, the region around the core which if quickly re-perfused may recover its function [75]. Lactate formed in the initial period of ischemia could remain in necrotic tissue and leave the region of injury after cell lysis in a period of weeks or months after the stroke onset.
Unlike to the increase of lactate, NAA is observed to slowly decrease over a time scale of hours after the induction of ischemia [75]. Several studies have described an initial rapid decrease in NAA of about 10% within the first few minutes followed by a slower decrease. It has been suggested that NAA diminish may be due to NAA degradation by enzymes within the injured neurons in the first few days or hours following infarction, or perhaps due to changes in other molecules (e.g. Glu, Gln, GABA etc.) which overlap with the spectral resonance of NAA [74].
tCho has been observed to either increase or decrease both in acute and chronic human ischemia [76]. Increases in Cho in stroke may be the result of gliosis or ischemic damage to myelin, while decreases are probably the result of edema, necrosis and cell loss [4]. Initial reduction in Cr/PCr is identified following infarction and further reductions have been demonstrated up to ten days following the time of onset [74]. Muniz Maniega et al. reported continuous reduction of Cr levels over a period of three months from the stroke onset [76].
A study by Rumpel et al. revealed that mI might also significantly contribute to the understanding of brain tissue response to ischemia, which is in line with a persistent cytotoxic swelling, attributed to the glial population, found in early subacute ischemic infarcts [77]. Acute ischemic also causes changes in the glutathione (GSH) system (decreased GSH in ischemic patients) as stroke is associated with significant oxidative stress [41]. Experimental studies have suggested that ischemic stroke may cause an increment in extracellular level of GABA; however there is very little work on the detection of GABA and glutamate in cerebral ischemia [78].
The term epilepsy covers a wide group of syndromes with varied etiology and prognosis. By providing an insight into the biochemical processes related to epileptic seizures, 1H-MRS aids in the localization or lateralization of the epileptogenic foci and in the influence of the metabolites concentration after the administration of antiepileptic drugs and/or after resection of the epileptogenic tissue.
Temporal lobe epilepsy (TLE) associated with hippocampal sclerosis (HS) is the most common refractory focal epilepsy. The localization is performed by the comparison of metabolites on the left and right temporal lobe, especially in the hippocampus and temporal poles, to determine which hemisphere is responsible for the genesis of seizures [79]. The metabolites of interest in epilepsy are NAA, GABA and glutamine/glutamate (Glx) and the less prominent mI and lactate (Figure 5). Most of the studies dealing with mesial TLE, demonstrate decreased levels of NAA in the affected temporal lobe when compared with controls or with the homologous non-epileptic contralateral region, with no changes or mild increases of tCho. Interestingly, not only decrease of NAA content occurs in the epileptogenic foci, but also unilateral presence of lactate in the mesian temporal lobe could potentially be indicative of the side of the epileptogenic zone [16].
Nowadays, a hypothesis exists in which the raise of mitochondrial energy consumption promotes a reduction of neuronal synthesis of NAA, and, therefore, an increase of glutamate (its precursor) [80]. In epileptic patients, it seems to exist a disequilibrium of glutamine/glutamate (Glx) and GABA [81]. Therefore, spectroscopic measurements of Glx complex could yield spatial information on the epileptogenic zone. However, although there is some evidence that Glx is elevated in TLE, its value as a marker for the epileptogenic zone has not been established yet.
Additionally, 1H-MRS studies of TLE have also been focused on mI, however, its role remains controversial. The study of Wellard et al.revealed elevated mI in the epileptogenic temporal lobe of patients with HS [82].They also reported a difference of mI levels between the seizure focus (temporal lobe) where mI is increased, and areas of seizure spread (frontal lobe) where mI is decreased. Thus, 1H-MRS may aid to the distinction of primary epileptogenic brain damage from seizure secondary effects on adjacent normal brain and help to distinguish drug refractory TLE patients, who will benefit from surgery by predicting postoperative outcome [83].
Short TE (35msec) spectra at 3T obtained in the left and right hippocampal formation from a patient with right HS using single-voxel technique. The decreased NAA signal and the increased mI at the affected region (A) are evident when compared with the contralateral normal hippocampal formation (B). Note the mild elevation of the Glx complex at the affected region (B).
Numerous studies have attempted to identify specific metabolic markers for different neurodegenerative diseases, such as Alzheimer’s dementia (AD) and Parkinson’s disease (PD), which concern loss of structure or function of neurons including death of the neuronal cells. The clinical objective in that cases is to establish a precise and early diagnosis as well as to understand the related brain changes that could help to slow down the course of the disease [60].
1H-MRS has been demonstrated to be highly specific and sensitive to the diagnosis of Alzheimer’s dementia (AD) [15]. Reduction in NAA is the most frequent 1H-MRS finding in AD [84, 85]. Single Voxel 1H-MRS studies have consistently found reductions in NAA/Cr ratio in the hippocampal formation [85] as well as other temporal regions [86] and the posterior cingulate gyrus [87]. Findings of reduced NAA have been also detected in the temporoparietal area, and the occipital lobes [85, 87]. There have been conflicting reports regarding Cho levels in patients with AD. Some researchers found elevated Cho and/or Cho/Cr ratios in AD patients, while others not [32, 85, 87]. Increased mI has been also observed, most often, in the temporal-parietal area [85], the posterior cingulate gyrus [32, 86, 87], the parietal white matter 86] and less often in the frontal lobes [86]. Few studies have also reported reduced Glx levels in AD patients compared to control subjects in the posterior cingulated gyrus [88] and lateral temporal cortex [86].
The majority of 1H-MRS studies in Parkinson’s Disease (PD) to date have primarily targeted brain levels of NAA, Cr, and Cho [89, 90]. Many researchers disclosed a significant reduction of ratios NAA/Cr and NAA/Cho in the temporoparietal cortex [91], the substantia nigra, the basal ganglia [92], the striatum or the occipital lobe [93]. Griffith et al. have demonstrated lower NAA/Cr ratios in the posterior cingulate gyrus of demented versus non demented subjects with PD [94]. Other investigators, however, have not detected such changes [95] in NAA, Cr, and Cho measurements, and the reasons for these different findings need to be resolved.
Gliomas are spatially heterogeneous lesions which arise from the ‘gluey’, or supportive tissue of the brain. The main types of gliomas are astrocytomas, oligodendrogliomas, and ependymomas. 1H-MRS is increasingly used in clinical studies to non-invasively identify regions with metabolic specific characteristics that reflect glioma type and grade.
A common observation in 1H-MRS of all glial tumors is a decreased levels of NAA and increased levels of tCho with a significant overlap among different glioma types [2, 22]. Thus, 1H-MRS is currently used primarily to differentiate glial tumor grade rather than to confirm a histopathological diagnosis [96].
However, the signal intensity of glutamine and glutamate (Glx) may aid the distinction between oligodendrogliomas and astrocytomas. Rijpkema et al. found significantly increased Glx levels for oligodendrogliomas when compared to that of astrocytomas [97] using short TE 1H-MRS. Additionally, in a study by Majos et al, ependymomas differentiated well from the other glial tumors by showing prominent peaks of mI+Gly and Taurine at long TE spectra [98].
Discrimination between tumor grades in gliomas is an important clinical issue, because there is a dispute on the optimum treatment strategy for patients with low-grade tumors. It remains an open question whether 1H-MRS is able to define WHO grade of gliomas. However, a recent study by Porto et al. revealed a more prominent loss of NAA and increase of tCho in WHO III over WHO II astrocytomas [99]. They consequently proposed NAA/tCho ratio as the most accurate index to discriminate between those tumor grades which is in agreement with what it is generally accepted, i.e. NAA/tCho ratios decrease with higher histological grade of gliomas.Law et al.demonstrated a threshold value of 1.6 for tCho/NAA which provided 74.2% sensitivity and 62.5% specificity in predicting the presence of a high-grade glioma [100]. Thus it is obvious that there is a consistent correlation between Cho increase as well as NAA decrease and tumor grade.
A study by Moller-Hartmann et al. revealed that instead of tCho, the amount of lipids proved to be the second-best discriminator between low- and high-grade gliomas, with glioblastomas multiforme (GBM) to exhibit the highest amount of lipids since necrosis is one of their microsopic hallmarks [3]. Although it has been previously proved that lactate also increases with grade, it is not always significantly differentiated between high and low grade gliomas [22]. Poor correlation between tumor grade and lactate is most likely due to the difficulty of accurately quantifying lactate in the presence of high lipid signals.
Short TE studies have also shown that mI levels may aid tumor classification and grading [22, 25]. Specifically, Castillo et al. retrospectively studied 34 patients with astrocytomas and found a trend towards lower mI levels in high-grade compared with low-grade tumors [25].
One of the most interesting results of the study by Server et al. was the elevation in the peritumoral Cho/Cr and Cho/NAA metabolite ratios in relation to glioma grading [101]. Thereby, as gliomas are infiltrating intracerebral tumors, 1H-MRS may allow to readily appreciate their grade in the perifocal region.
Cerebral metastases are a common complication of cancer and can affect 20% to 40% of patients [102] who suffer from primary tumors in lung, breast, skin or colon.
When a metastatic brain tumor presents as a solitary lesion, it is usually indistinguishable from a high grade glioma [103]. Their distinction is important because the treatment approach and follow-up are different for these two different tumors.
The potential of in vivo 1H-MRS for differentiating intracerebral metastases from GBMs has been investigated in a number of studies [102, 104]. Older studies [22, 105] have reported that intratumoral 1H-MRS, either on short or long TE, was unable to differentiate between metastases and GBMs, as they share common metabolic features. Those concern increased levels of lipids and tCho and reduced levels of NAA as it is depicted in figure 6. Nevertheless, a study by Moller-Hartman et al. revealed elevated lipids for metastases, with statistically significant difference from GBMs [3]. Opstad et al. speculated that the differences in lipid profiles may be related to differences of membrane structures of infiltrative versus migratory tumor cells [106]. Significantly higher Cho/Cr ratio for metastases than for GBMs was reported by Server et al. due to GBMs higher levels of necrosis [102]. On the contrary, Law et al. revealed significantly lower Cho/Cr ratio for metastases than for high grade gliomas [105]. These conflicting results may be due to the intrinsic heterogeneity of such tumors.
Typical Short TE spectra from glioblastoma multiforme (A) and intracerebral metastases (B).
Short TE spectra from peritumoral areas of glioblastoma multiforme (A) and intracerebral metastases (B).
Furthermore, promising results in differentiating between GBMs and metastases by means of the resonances at 3.56 ppm, represented by the sum of mI and Gly, have been previously observed [36]. Gly/mI showed a tendency to be higher in GBMs than in metastases.
Measuring the peritumoral metabolites or metabolic ratios is often more useful in differentiating intracranial metastases from high grade gliomas with more reproducible results among different studies. Elevated Cho as well as reduced NAA have been found in the peritumoral region of high-grade gliomas, but not in the peritumoral region of metastases when compared to normal levels as it is illustrated in figure 7.
Those findings support the hypothesis that the edema surrounding metastases is purely vasogenic, while the peritumoral region of GBMs is characterized by extensive infiltration of tumor cells [102, 104-106].
Some patients get their brain metastases detected before the primary cancer. Since GBM case has being withdrawn from the differential diagnosis, identification of metastases type would be important for further treatment. Sjobakk et al. investigated the feasibility of using 1H-MRS to characterize brain metastases originating from different primary cancers. The results presented in their study, demonstrated that lipid signals on both short and long TE spectra are important for metastases characterization. Although non-statistically significant, lung metastases tended to differentiated from breast metastases in respect to their lipid signals, while the melanoma showed no trend [107]. Chernov et al retrospectively studied 25 metastatic brain tumors from lungs, colon, breast, kidney, prostate and cardiac muscle, using 1H-MRS on long TE. The detected metabolic characteristics revealed that metastases of colorectal carcinoma have significantly greater lipid content, expressed as Lipids/Cr ratio, compared to metastatic tumors of other origin. The authors suggested an optimal Lipids/Cr cut off value of 2 for the identification of the colorectal carcinoma [108]. It is obvious that 1H-MRS may aid in the determination of cerebral metastases origin, nevertheless, further research is needed to determine the exact role of proton MR spectroscopy in the identification of the tissue type of metastatic brain tumors.
Meningiomas are common intracranial tumors and are generally easily diagnosed by their characteristic radiological imaging appearance of solid mushroom imaging pattern, extracranial location, dura matter conjunction and sinus involvement. However, 15% of meningiomas exhibit rim like enhancement, a prominent cystic component, hemorrhage, or even metaplasia [109], mimicing gliomas or cerebral metastatic tumors. 1H-MRS has been proved useful in differentiating meningiomas with atypical radiologic pattern from other brain tumors [36].
Alanine at 1.47ppm has been considered as the characteristic metabolic marker of meningiomas which differentiates them from other brain tumors [36, 38]. Nevertheless, reported occurrence of Alanine varies among different studies [37, 16] as it can significantly overlap with lactate resonance due to J-coupling effect [37].
In the absence of Alanine, several investigators aimed to correlate other metabolites to meningioma presence. Studying the metabolic profile of different cerebral tumors using short TE 1H-MRS, Howe et al. found that low levels of mI and Cr were characteristic for meningiomas relative to grade II astrocytomas, anaplastic astrocytomas and glioblastomas [22]. In the same study meningiomas revealed the highest Cho/Cr ratio among the other brain tumors, on both short and long TE. Another reported specific finding for meningiomas, is the absence of the neuronal marker NAA. Instead of partial volume effects [3], the peak of NAA at meningioma spectra, may also represent other endogenous NAA compounds (NACs) such as N-acetylaspartylglutamate, N-acetylneuraminic acid and N-acetylgalactosamine [37].
A recent study by Kousi et al.revealed a distinct chemical compound, observed in all meningiomas recruited for that study, which may establish a rather specific marker in their differential diagnosis from high grade gliomas and metastases [6]. This chemical substance, resonated at 3.8ppm using short TE 1H-MRS (Figure 8) and according to the in vitro study of Tugnoli et al. it might receive contribution from phosphoethanolamine (PE) and other amino acids such as Leukine, ALanine, Glutamate, Glutamine, Glutathione, Lysine, Arginine and Serine [110].
FLAIR T2 images (left) of a meningioma with its corresponding spectrum (right),at short TE (35ms). Elevated Cho and lipid resonances were detected at 3.2ppm and 1.3ppm respectively, as well as a distinct chemical compound resonating at 3.8 ppm [6].
Primary central nervous system lymphoma (PCNSL) represents1% of all brain tumors and its incidence has increased in the last 3 decades.Although densely contrast-enhancing lesions, without the presence of necrosis are characteristic imaging features of PCNS lymphoma, it can be difficult, sometimes even impossible, to distinguish PCNSLs from high grade gliomas on conventional MRI [111]. Their differentiation, however, has important diagnostic and therapeutic implications.
For the correct diagnosis of brain lymphomas, 1H-MRS has reported promising results. The most specific finding for PCNSL on MRS is an increase in lipid and Cho resonances (Figure 9). Sometimes, lipid peaks inPCNSL may be more prominent than in high gradegliomas and can help differentiate between the twotumor types [107].
Lipids are typically a signature of cell death; however, a lipid dominated spectrum found in PCNSL does not indicate necrosis.This appears to be due to numerous macrophages and the increased turnover of membrane components in transformed lymphoid cells which contain high concentrations of mobile lipids [112].
Histopathologically, PCNSLs are characterized by a diffusely infiltrative pattern and hence, it is important to survey the peritumoral area also and not just the area of obvious tumor involvement. Like high grade gliomas, the peritumoral area of PCNSLs demonstrates an abnormal metabolite pattern. Chawla et al. reported increased Cho/Cr and Lip+Lac/Cr ratios in the peritumoral area of PCNSLs. They also observed significantly higher Lip+Lac/Cr ratio in the peritumoral area of PCNSLs when compared with that of GBMs, suggesting the presence of infiltrative active lymphocytes and macrophages in areas beyond lymphoma boundaries. Using a threshold value of 7.09 for Lip+Lac/Cr ratio they differentiated PCNSLs from GBMs with 84.6% sensitivity and 75% specificity [107]. Therefore, in the absence of obvious necrosis, increased lipid concentration together with a markedly elevated Cho/Cr ratio for both intratumoral and peritumoral areas can provide important metabolic information which may improve the distinction between PCNSLs and other brain tumors.
Spectra from an intracerebral lymphoma on both short (A) and long TE (B), demonstrating the characteristic elevation of lipid and Cho resonances.
Central neurocytomas (CNCs) are a neuronal tumor almost exclusively located in the lateral ventricles that appear in young adults. Most of these tumors do not recur after surgery and are generally considered benign, with a favorable prognosis [28].
Instead of their typical radiological appearance - a well circumscribed lobulated mass in the anterior portion of the lateral ventricles– it may not always be possible to differentiate CNCs from other intraventricular tumors such as oligodendrogliomas and ependymomas [113].
1H-MRS has been proved a valuable tool for the presurgical diagnosis of these neoplasms. Previous studies have reported CNCs to consistently show the tumoral pattern of increased Cho and decreased NAA levels [28, 35]. On the contrary, lactate has not been observed consistently in all studies. Specifically, although Kim et al. reported lactate in all of their patients, Shah et al. observed lactate in only 9% of the CNC cases [114, 115]. A few studies have speculated the rather specific marker of Gly at 3.55ppm on long TE spectra, strongly suggesting CNC occurrence [28, 115].
The presence of Ala in a patient with CNC was first reported by Chuang et al. using a 3T MR system [116]. It was demonstrated as an inverted doublet at 1.5 ppm with a TE of 135 msec. Similarly, Krishnamoorthy et al. also observed Ala in all three CNC cases (100%) studied, whereas in the study of Shah et al. Ala was observed in 64% of the CNCs [35, 115].
Thus, CNCs may show Ala as an inverted doublet at 1.5 ppm in long TE spectra. Although one may observe Ala in other intraventricular tumors such as meningioma, other characteristic peaks such as Gly, high Cho and decreased NAA should help to correclty identify CNC.
Gliomatosis Cerebri (GC) is a rare brain tumor characterized by a diffuse neoplastic overgrowth of glial elements of various histological subtypes (astrocytoma, oligodendroglioma, or mixed glioma) and extensive infiltration of at least two lobes [117]. Unlike gliomas, the neuronal architecture is usually preserved [118].
MRI characteristics of GC are non-specific and occasionally it is difficult to differentiate GC from demyelinating diseases or viral encephalitis, and biopsy is often inconclusive [9].Given the unfavorable prognosis of this tumor type, there is a demand for alternative imaging techniques, such as 1H-MRS, to grade GC and to detect the most anaplastic areas for determining surgical areas and radiotherapeutic targets.
A few studies have looked at the spectral features of such tumors and those are consistent with the spectral features of gliomas discussed above. By studying 8 patients with GC using long TE 1H-MRS, Bendszus et al. found elevated Cho/Cr and Cho/NAA ratios, as well as decreased NAA/Cr ratios of varying degrees in the abnormal areas on T2-weighted images [119]. Similarly, a retrospective analysis by Yu et al. also revealed high Cho/Cr and Cho/NAA ratios and low NAA/Cr ratio within the areas of hyperintensity on T2-weighted images in 8 histopathologically confirmed patients with GC. Anaplastic areas had higher Cho/NAA ratio and the lactate doublet was present [117]. Apart from being beneficial in the grading of GC, 1H-MRS might reflect the true extent of neoplastic infiltration more accurately than MRI. Bendszus et al. found elevated Cho/Cr and Cho/NAA ratios in the tumor margins that appeared normal on T2- weighted images. Tumoral infiltration of the margin of the lesion that appeared normal on T2-weighted images was also confirmed by Yu et al. by observing increased Cho/NAA ratio in those areas [117].
From the aforementioned findings it can be concluded that 1H-MRS when combined to MRI findings may aid to GC diagnosis. Additionally, the determination of highly anaplastic areas and areas of tumoral infiltration may have a great impact in radiotherapy planning.
Pituitary adenomas and craniopharyngiomas, are the most frequent suprasellar space occupying lesions and are generally regarded as benign neoplasms of the pituitary gland. Nevertheless, with respect to the differential diagnosis of suprasellar masses, pituitary adenomas, craniopharyngiomas together with gliomas and meningiomas can be considered [120].
To date only a few cases of pituitary adenomas and craniopharyngiomas have been studied by in-vivo 1H-MRS [16, 120, 121], probably because of their relative rarity and the technical difficulties in obtaining in vivo high-quality spectra without artifacts is such a region [120]. In a study by Chernov et al., the vast majority of the 19 pituitary adenomas were characterized by a significant reduction of NAA peak, moderate elevation of Cho, and infrequent presence of small lipid and lactate peaks. This metabolic pattern differentiated them from low grade gliomas which showed a moderate decrease of NAA and Cr peaks. In the same study, craniopharyngiomas were typically characterized by a significant decrease of all metabolites and presence of multiple additional peaks which were possibly resulted from the presence of calcifications and microcysts within the investigated volume of tissue [16]. On the contrary, Sener et al. demonstrated very prominent peaks in the craniopharyngiomas between 0.5 and 1.5 ppm, which probably corresponded to lipid peaks. Histological findings also revealed high amounts of cholesterol, lipids and lactate in the cyst fluid correlating with their spectroscopic findings [120].
1H-MRScan provide important in vivo metabolic information, complementing morphological findings from conventional MRI in the clinical setting. This technique is an extremely valuable tool in solving difficult neurological cases and increase confidence in diagnosis; however, it should be always considered a supplementary tool to the patients’ clinical history, examination, and conventional MRI when reaching the final diagnosis.
The future would be to combine 1H-MRS with other advanced magnetic resonance techniques such as Diffusion/Diffusion Tensor Imaging and Perfusion-weighted Imaging, which will potentially prove to be useful in both clinical and research settings. Ultimately, these advanced tools may be used in a multiparametric, algorithmic fashion to characterize tissue biology and dramatically improve tumor differential diagnosis.
Radiation oncology is the discipline dealing with the treatment of malignant neoplasias or cancerous lesions (and occasionally benign lesions) with ionizing radiation for cure or palliation intent. The clinical modality or technique has been used to treat the patient in radiation oncology is referred to as radiation therapy (or “radiotherapy”). Radiotherapy has often given in combination with other treatment modalities for instance chemotherapy, surgery, hormonal therapy, etc. The aim of radiotherapy is to deliver a precisely measured dose of irradiation to a defined tumor volume with as minimal damage as possible to surrounding healthy tissue, resulting in eradication the tumor, high quality of life, and prolongation of survival [1]. Figure 1 presents a typical radiotherapy workflow, from patient consult and assessment to follow-up. The field of radiotherapy has witnessed with significant technological advances over the last decades. This advancing has introduced the complexity of radiotherapy processes and generating a massive amount of data (also so-called “big data”) during radiotherapy workflow.
\nRadiotherapy workflow, from patient consult and assessment to follow-up.
Big data is data which is of a large volume, often combining multiple data sets and requiring innovative forms of information technology to process this data [3]. Big data has characterized by four V’s: volume, variety, velocity and veracity [3]. In radiation oncology, data can be categorized as “Big Data” because (a) the use of data-intensive imaging modalities (volume), (b) the imaging archives are growing rapidly (velocity), (c) there is an increasing amount of imaging and diagnostic modalities available (variety), and (d) interpretation and quality differs between care providers (veracity) [4]. The radiation oncologists are overwhelmed with scientific literature, rapidly evolving treatment techniques, and the exponentially increasing amount of clinical data [5]. Figure 2 shows more and more information is associated with the patient as the proceeds along the radiotherapy process, like a snowball rolling down a hill [2]. The radiation oncologists need help translating all these data into knowledge that supports decision-making in routine clinical practice [6, 7, 8, 9, 10].
\nWith each step along the radiotherapy workflow, more information is created and collected which has associated with the patient (reproduced from [2]).
In this direction, such collaborative efforts have been established in the last few years to advance the possibilities of using big data to facilitate personalized clinical patient care in the field of radiation oncology. For example, in 2015, the American Society for Therapeutic Radiation Oncology (ASTRO), National Cancer Institute (NCI), and American Association of Physicists in Medicine (AAPM) co-organized a workshop with aims focused on opportunities for radiation oncology in the era of big data [9]. Later in 2017, the American College of Radiology (ACR) has established the Data Science Institute (DSI) with a core purpose to empower the advancement, validation, and implementation of artificial intelligence (AI) in medical imaging and the radiological science for the benefit of patients, society, and the profession [10].
\nMachine learning (ML), a branch of artificial intelligence, is the technology of developing computer algorithms that are able to emulate human intelligence. An ML algorithm is a computational process that uses input data to achieve the desired task without being literally programmed (i.e., “hard-coded”) to produce a particular outcome [2]. These algorithms are in a sense “soft-coded” in that they automatically alter or adapt their architecture through repetition (i.e., experience) so that they become better and better at achieving the desired task [2]. The process of adaptation is called training, in which samples of input data have provided along with desired outcomes [2]. The algorithm then optimally configures itself so that it cannot only provide the desired result when presented with the training inputs, but it can even generalize to produce the desired outcome from new data [2]. Figure 3 shows a generic ML workflow. In which, the ML model is trained first on a training data then the trained model is used for predicting the results for new data [2]. More deeply, ML algorithms have been classified according to the nature of the data labeling into supervised (e.g., classification or regression), unsupervised (e.g., clustering and estimation of probability density function), and semi-supervised learning approach (e.g., text/image retrieval systems) [11, 12, 13].
\nA generic machine learning workflow.
With the era of big data, the utilization of machine learning algorithms in radiation oncology research is rapidly growing. Its applications include treatment response modeling, treatment planning, organ segmentation, image-guidance, motion tracking, quality assurance, and more. In this chapter, we provide the interested reader with an overview about the ongoing advances and cutting-edge applications of the ML methods in radiation oncology from a workflow perspective, from patient diagnosis and assessment to treatment delivery and follow-up. We present the areas where ML could be applied to improve the efficiency, i.e., optimizing and automating the clinical processes, and quality, i.e., potentials for decision-making support toward precision medicine in radiation therapy, of patient care. This chapter is organized as follows: Section 1 provides introduction to radiation oncology, big data, and machine learning concept; Section 2 illustrates an overview of the utilization of machine learning methods in radiation oncology research from a workflow perspective; Section 3 discusses limitations and the challenges of the of the current approaches as well as the future vision to overcome these problems; and Section 4 presents conclusions.
\nThe utilization of machine learning algorithms in radiation oncology research has covered almost every part in radiotherapy workflow process (Figure 1). ML techniques could compensate for human limitations in handling a large amount of flowing information in an efficient manner, in which simple errors can make the difference between life and death. Also, it would allow improvements in quality of patient care through the potentials toward a practical application of precision medicine in radiation oncology. In this section, we go over each part in the radiation oncology workflow (Figure 1) process presenting studies that have been conducted with machine learning models. The radiation oncology workflow starts with patient diagnosis and assessment, to treatment simulation, to treatment planning, to quality assurance and treatment delivery, to treatment outcome and follow-up.
\nThe radiation oncology process begins at the first consultation. During which, the radiation oncologist and patient meet to discuss the clinical situation to determine a treatment strategy [14]. The stage that precedes the patient assessment and consultation is a patient diagnosis, in which patient with cancer disease identified on medical images and then pathologically confirmed the disease. Machine learning toolkits such as computer-aided detection/diagnosis have been introduced for identifying and classifying cancer subtypes (staging). For example, lesion candidates into abnormal or normal (identify and mark suspicious areas in an image), lesions or non-lesions (help radiologists decide if a patient should have a biopsy or not), malignant or benign (report the likelihood that a lesion is malignant), etc. Machine learning plays a crucial role in computer-aided detection/diagnosis toolkits, and it could provide a “second opinion” in decision-making to the physician in diagnostic radiology.
\nComputer-aided detection (CADe) has defined as detection made by a physician/radiologist who takes into account the computer output as a “second opinion” [2]. CADe has been an active research area in medical imaging [2]. Its task is classification based solving a problem, in which the ML classifier task here is to determine “optimal” boundaries for separating classes in the multidimensional feature space. It focuses on a detection task, e.g., localization of lesions in medical images with the possibility of providing the likelihood of detection.
\nSeveral investigators [15, 16, 17, 18] have developed ML-based models for detection of cancer, e.g., lung nodules [15] in thoracic computed tomography (CT) using massive training artificial neural network (ANN), micro-calcification breast masses [16] in mammography using a convolutional neural network (CNN), prostate cancer [17] and brain lesion [18] on magnetic resonance imaging (MRI) data using deep learning. Chan et al. [16] achieved a very good accuracy, an area under a receiver operating characteristic curve (AUC) of 0.90, in the automatic detection of clustered of breast microcalcifications on mammograms. Suzuki et al. [15] reported an improved accuracy in the detection of lung nodules in low-dose CT images. Zhu et al. [17] reported an averaged detection rate of 89.90% of prostate cancer on MR images, with clear indication that the high-level features learned from the deep learning method can achieve better performance than the handcrafted features in detecting prostate cancer regions. Rezaei et al. [18] results demonstrated the superior ability of the deep learning approach in brain lesions detection.
\nOverall, the use of computer-aided detection systems as a “second opinion” tool in identifying the lesion regions in the images would significantly contribute to improving diagnostic performance. For example, it would lead to avoid missing cancer regions, increase sensitivity and specificity of detection (increased accuracy), and diminish inter- and intraobserver variability.
\nComputer-aided diagnosis (CADx) is a computerized procedure to provide a “second objective opinion” for the assistance of medical image interpretation and diagnosis [19]. Similar to CADe, its task is a classification solving-problem. CADx focuses on a diagnosis (characterization) task, e.g., distinction and automatically classifying a tumor or lesion being malignant or benign with a possibility of providing the likelihood of diagnosis.
\nNumerous studies [19, 20, 21, 22] have demonstrated the application of CADx tools for diagnosing lung [19, 20, 21] and breast [19, 22] lesions. Cheng et al. [19] investigated the deep learning capability for the diagnosis of breast lesions in ultrasound (US) images and pulmonary nodules in CT scans. Their results showed that the deep-learning-based CADx can achieve better differentiation performance than the comparison methods across different modalities and diseases. Figure 4 illustrates several cases of breast lesions and pulmonary nodules in US and CT images, respectively, differentiated with deep learning-based CADx [19]. Feng et al. [20] and Beig et al. [21] studied the classification of lung lesions on endo-bronchoscopic images [20] with logistic regressions, and non-small cell lung cancer (NSCLC) adenocarcinomas distinctions from granulomas on non-contrast CT [21] using support vector machine (SVM) and neural network (NN). The reported results indicated an accuracy of 86% in distinguishing lung cancer types, e.g., adenocarcinoma and squamous cell carcinoma [20]. Surprisingly, the reported results [21] in distinguishing non-small cell lung cancer adenocarcinomas from granulomas on non-contrast CT images showed that the developed CADx systems outperformed the radiologist readers. Joo et al. [22] developed a CADx system using an ANN for breast nodule malignancy diagnosis in US images. Their results demonstrated the potential to increase the specificity of US for characterization of breast lesions.
\nComputer-aided diagnosis for lung nodules and breast lesion with deep learning. It shows that it may be hard to differentiate for a person without a medical background and for a junior medical doctor (reproduced from [19]).
Overall, computer-aided diagnosis tool as a “second opinion” system could significantly enhance the radiologists’ performance by reducing the misdiagnosed rate of malignant cases, then decreases the false positive of the cases sent for surgical biopsy. Also with CADx, the diagnosis can be performed based on multimodality medical images in a non-invasive (without biopsy), fast (fast scanning) and a low-cost way (no additional examination cost).
\nDuring the patient assessment phase, the radiation oncologist and patient meet to discuss the clinical situation. Circumstances like the risks and benefits of treatment and the patient’s goals of care are determined for the treatment strategy [14]. Useful information to assess the potential benefit of treatment is acquired, e.g., tumor stage, prior and current therapies, margin status if post-resection, ability to tolerate multimodality therapy, and overall performance status [14]. Parameters that impact potential risk and tolerability of treatment are balanced, e.g., patient age, comorbidities, functional status, the proximity between tumor and critical normal tissues, and ability to cooperate with motion management [14]. All of these represent valuable features which can be utilized to build predictive models of treatment outcome and toxicity. These models, then, can be used to inform physicians and patients to manage expectations and guide trade-offs between risks and benefit [14].
\nMachine learning models [23, 24, 25, 26] such as logistic regressions, decision trees, random forests, gradient boosting, and support vector machines are suitable for this purpose. Logistic regressions or decision trees are similarly effective [23, 24] for a goal to assist physicians and patients reach the best decision, compromising balance between interpretability of the results and accurate predictions. In case of accuracy is favored over interpretability, then methods [25, 26] such as random forests or gradient boosting, and SVMs with kernels, are better and consistently win most modeling competitions [14].
\nOverall, the delivery of models that could help with these scenarios require standardizing nomenclature and developing standards for data collection of these heterogeneous patient clinical data remain a challenge in radiation oncology.
\nOnce a physician and patient have decided to proceed with radiation therapy, the physician will place robust instructions for a simulation, which is then scheduled. The order for simulation includes details about immobilization, scan range, treatment site, and other specifics necessary to complete the procedure appropriately [14]. Patient preparation for simulation could include fiducial placement, fasting or bladder/rectal filling instructions, or kidney function testing for intravenous (IV) contrast. Special instructions have given for patients with a cardiac device, or who are pregnant, and lift help or a translator is requested if necessary [14]. The treatment simulation process typically includes patient’s setup and immobilization, three- or four-dimensional computed tomography (3DCT or 4DCT) image data acquisition, and image reconstruction/segmentation. Machine learning algorithms could have an essential role to play in this sequence to improve the simulation quality, hence a better treatment outcome.
\nThree-dimensional CT anatomical image information for the patient are acquired during the simulation on a dedicated CT scanner (“CT-Simulator”) to be used later for the treatment planning purposes. A good CT simulation is critical to the success of all subsequent processes, to achieve an accurate, high quality, robust, and deliverable plan for a patient. It could prevent a repeated CT simulation due to insufficient scan range, suboptimal immobilization, non-optimal bladder/rectal filling, artifacts, lack of breath-hold reproducibility, and so on [14]. 4DCT scanning is used increasingly in radiotherapy departments to track the motion of tumors in relation to the respiratory cycle of the patient. It monitors the breathing cycle of the patient and can either; acquire CT images at a certain point in the breathing cycle, or acquire CT images over the whole breathing cycle. This CT data is then used to generate an ITV (internal target volume) that encompasses the motion of the CTV (clinical target volume), or MIP (maximum intensity projection) scans to aid in the definition of an ITV [2]. 4DCT imaging is necessary for successful implementation of stereotactic ablative radiotherapy (SBRT), e.g., for early-stage NSCLC.
\nFew works [27, 28, 29, 30] have carried out using ML-based methods for this purpose. For instance, a work by Fayad et al. [27] demonstrated an ML method based on the principal component analysis (PCA) to develop a global respiratory motion model capable of relating external patient surface motion to internal structure motion without the need for a patient-specific 4DCT acquisition. Its finding looks promising but future works of assessing the model extensively are needed. Another study by Steiner et al. [28] investigated an ML-based model on correlations and linear regressions for quantifying whether 4DCT or 4D CBCT (cone-beam CT) represents the actual motion range during treatment using Calypso (Varian Medical Systems Inc., Palo Alto, CA, USA) motion signals as the “ground truth.” The study results found that 4DCT and 4DCBCT under-predict intra-fraction lung target motion during radiotherapy. A third interesting one by Dick et al. [29] examined an ANN model for fiducial-less tracking for the radiotherapy of liver tumors through tracking lung-diaphragm border. The findings showed that the diaphragm and tracking volumes are closely related, and the method has indicated the potential to replace fiducial markers for clinical application. Finally, a study by Johansson et al. [30] investigated an ML-based PCA model for reconstructing breathing-compensated images showing the phases of gastrointestinal (GI) motion. Its results indicated that GI 4D MRIs could help define internal target volumes for treatment planning or support GI motion tracking during irradiation.
\nOverall, the discussed ML-based methods in the simulation area have shown the potential for improved accuracy of patient CT simulation. Machine learning utilization in 3D/4D CT image acquisition simulation is an area where the community has focused little effort. Thus, focusing on the simulation, there are many questions that could be answered/optimized through ML algorithms to aid in decision-making and overall workflow efficiency.
\nHere, we explore the power of machine learning based methods for image reconstruction in radiation oncology procedure. We present two application examples where ML has utilized for estimating CT from MRI images and reconstructing a 7 Tesla (7 T)-like MR image from a 3 T MR image.
\nThe first application supports reconstructing an image modality form another imaging modality, e.g., CT image from MR image. Clinical implementation of MRI-only treatment planning radiotherapy approach requires a method to derive or reconstruct synthetic CT image from MR image. CT is currently supporting the workflows of radiation oncology treatment planning for dose calculations. However, CT imaging modality has some limitations in comparison with other modalities like MRI, e.g., (a) CT images provide poor soft tissue contrast compared to MRI scans which has superior visualization of anatomical structures and tumors, and (b) CT exposes radiation during CT imaging, which may cause side effect to the patient, where MRI is much safer and does not involve radiation.
\nNumerous studies [31, 32, 33, 34] have demonstrated ML-based approaches to map CT images to MR images like deep learning (fully CNN) model [31], boosting-based sampling (RUSBoost) algorithm [32], random forest and auto-context model [33], and U-net CNN model [34]. Nie et al. [31] experimental results showed that deep learning method is accurate and robust for predicting CT image from MRI image. Figure 5 shows the synthetic CT image from MRI data with deep learning and the “ground truth” MRI [31]. The developed deep learning model outperformed other state-of-the-art methods under comparison. Bayisa et al. [32] proposed an approach based on boosting algorithm indicated outperformance in CT estimation quality in comparison with the existing model-based methods on the brain and bone tissues. Huynh et al. [33] experimental results showed that a structured random forest and auto-context based model can accurately predict CT images in various scenarios, and also outperformed two state-of-the-art methods. Chen et al. [34] investigated the feasibility of a deep CNN for MRI-based synthetic CT generation. The gamma analysis of their results with “ground truth” CT image for 1%/1 mm gamma pass rates was over 98.03%. The dosimetric accuracy on the dose-volume histogram (DVH) parameters discrepancy was less than 0.87% and the maximum point dose discrepancy within PTV (planning target volume) was less than 1.01% respect to the prescription on prostate intensity modulated radiotherapy (IMRT) planning.
\nSynthetic CT image from MRI data. MR image (left), estimated CT form the MR (middle) with deep learning, and “ground truth” (right) MR image for the same subject (reproduced from [31]).
Overall, the presented findings have obviously demonstrated the potential of the discussed methods to generate synthetic CT images to support the MR-only workflow of radiotherapy treatment planning and image guidance.
\nThe second application supports reconstructing a high-quality image modality from a lower quality one, e.g., 7 T-like MR image from 3 T MR image. The advanced ultra—high 7 T magnetic field scanners provide MR images with higher resolution and better tissue contrast compared to routine 3 T MRI scanners. However, 7 T MRI scanners are currently more expensive, less available in clinical centers, and higher restrictions are required for safety due to its extremely high magnetic field power. As a result, generating/reconstructing a 7 T-like MR image from a 3 T MR image with ML-based approaches would resolve these concerns as well as facilitate early disease diagnosis.
\nResearchers [35, 36, 37, 38] have developed ML-based models to generate a 7 T-like MR image from 3 T MR image. Approaches based on deep learning CNN [35], hierarchical reconstruction based on group sparsity in a novel multi-level canonical correlation analysis (CCA) space [36], and random forest and sparse representation [37, 38] have been investigated to map 3 T MR images to be as 7 T-like MR images. Bahrami et al. [35] visual and numerical results showed that deep learning method outperformed the comparison methods. Figure 6 presents the reconstruction of 7 T-like MR image from 3 T MR image with deep learning. A second study [36] done by the same author showed that a hierarchical reconstruction based on group sparsity method outperformed other previous methods and resulted in higher accuracy in the segmentation of brain structures, compared to segmentation of 3 T MR images. Other studies by Bahrami et al. [37, 38] using random forest regression model and a group sparse representation showed that the predicted 7 T-like MR images can best match the “ground-truth” 7 T MR images, compared to other methods. Moreover, the experiment on brain tissue segmentation showed that predicted 7 T-like MR images lead to the highest accuracy in the segmentation, compared to segmentation of 3 T MR images.
\nReconstruction of 7 T-like MR image from 3 T MR image. 3 T MR image (left), reconstructed 7 T-like MR image (middle) using deep learning, and 7 T MR “ground truth” image (left) of the same subject with each one corresponded with a same selected zoomed area. From the figure, 7 T MR image shows clearly better anatomical details and tissue contrast compared to 3 T MR image (reproduced from [35]).
Overall, the predicted 7 T-like MR images have demonstrated better spatial resolution compared to 3 T MR images. Moreover, delineation critical structure, i.e., brain tissue structures on 7 T-like MR images showed better accuracy compared to segmentation of 3 T MR images. Adding to above, such high-quality 7 T-like MR image could better help disease diagnosis and intervention.
\nImage registration in radiotherapy is the process of aligning images rigidly which allows some changes in images to be easily detected. However, such an alignment does not model changes from, e.g., organ deformation, patient weight loss, or tumor shrinkage. It is possible to take such changes into account using deformable image registration (DIR) which is a method for finding the mapping between points in one image and the corresponding points in another image. DIR has the perspective of being widely integrated into many different steps of the radiotherapy process. The tasks of planning, delivery, and evaluation of radiotherapy can all be improved by taking organ deformation into account. Use of image registration in image-guided radiotherapy (IGRT) can be split into intra-patient (inter- and intra-fractionated) and inter-patient registration. Intra-patient registration is matching of images of a single patient, e.g., inter-fractional registration (i.e., improving patient positioning, and evaluating organ motion relative to bones) and intra-fractional registration (i.e., online tracking of organ movement). In contrast, inter-patient registration is matching images from different patients (i.e., an “average” of images acquired from a number of patients, thereby allowing information to be transferred from the atlas to the newly acquired image). The process of combining information from two images after these have been registered is called data fusion. A particular use of data transfer between images is the propagation of contours from the planning image or an atlas to a newly acquired image [39, 40]. Although many image registration methods have been proposed, there are still some challenges for DIR of complex situations, e.g., large anatomical changes and dynamic appearance changes. Advancement in computer vision and deep learning could provide solutions to overcome these challenges of conventional rigid/deformable image registrations.
\nVarious machine learning-based methods [41, 42, 43, 44, 45, 46, 47] for image registration have proposed by investigators to not only align the anatomical structures but also alleviate the appearance difference. Hu et al. [41] proposed a method based on regression forest for image registration of two arbitrary MR images. The learning-based registration method achieved higher registration accuracy compared with other counterpart registration methods. Zagoruyko et al. [42] proposed a general similarity function for comparing image patches, which is a task for many computer vision problems. The results showed that such an approach like CNN-based model can significantly outperform other state-of-the-art methods. Jiang et al. [43] employed a discriminative local derivative pattern method to achieve fast and robust multimodal image registration. The results revealed that the proposed method can achieve superior performance regarding accuracy in multimodal image registration as well as also indicated the potential for clinical US-guided intervention. Neylon et al. [44] developed a deep neural network for automated quantification of DIR performance. Their results showed a correlation between the NN predicted error and the “ground truth” for the PTV and the organs at risk (OARs) were consistently observed to be greater than 0.90. Wu et al. [45, 46] developed an NN-based registration quality evaluator, and a deep learning-based image registration framework, respectively, to improve the image registration robustness. The quality evaluator method [45] showed potentials to be used in a 2D/3D rigid image registration system to improve the overall robustness, and the new image registration framework [46] consistently demonstrated more accurate registration results when compared to the state-of-the-art. Kearney et al. [47] developed a deep unsupervised learning strategy for CBCT to CT deformable image registration. The results indicated that deep learning method performed better than rigid registration, intensity corrected demons and landmark-guided deformable image registration for all evaluation metrics.
\nOverall, most of the machine learning based methods discussed here for image registration have revealed superior performance regarding accuracy in multimodal image registration. Hence, potentials for improved rigid/deformable image registration in radiation oncology are clinically feasible.
\nVolume definition is a prerequisite for meaningful 3D treatment planning and for accurate dose reporting. International Commission on Radiation Units and Measurements (ICRU) Reports No. 50, 62, 71 and 83 [48] define and describe target volumes (e.g., planning target volume) and critical structure/normal tissue (organ at risk) volumes that aid in the treatment planning process and that provide a basis for comparison of treatment outcomes. The organ at risk is an organ whose sensitivity to radiation is such that the dose received from a treatment plan may be significant compared with its tolerance, possibly needs to be delineated to evaluate its received dose [49]. Multimodal diagnostic images, e.g., CT, MRI, US, positron emission tomography (PET)/CT, etc. can be used through image fusion to help in the process of delineating tumor and OAR structures on CT slices acquired during the patient’s treatment simulation. The delineation (auto-contouring) process has subsequently become performed via automated or semi-automated analytical model-based software commercially available for clinical use (e.g., Atlas based-models). These software tools are performing reasonably well for critical organs/OARs delineation but not yet ready for tumor/target structures contouring which represent a challenging task. State-of-the-art machine learning algorithms may play an effective role here for both tasks.
\nSeveral ML-based methods [52, 53, 54, 55, 56, 57, 58] have reported for tumor/target segmentation/auto-contouring, e.g., brain [52, 53, 54, 55], prostate [56], rectum [57], sclerosis lesion [58], etc. The reported results showed that deep learning [54, 55] and ensemble learning [50, 53] ML-based methods are the winner algorithms over the other ML-based methods in the brain tumor segmentation competitions [50]. Such a method by Osman [52] based on SVM for glioma brain tumor segmentation showed a robust consistency performance on the training and new “unseen” testing data even though its reported accuracy on multi-institution datasets was reasonably acceptable. Figure 7 shows the whole glioma brain tumor segmentation on MRI (BRATS’2017 dataset [50, 51]) with an SVM model [52]. For organs segmentation, deep learning algorithm [57, 59, 60] has shown a superior performance than other state-of-the-art segmentation methods and commercially available software for segmentation of, e.g., rectum [57], parotid [59], etc.
\nWhole glioma brain tumor segmentation on MRI (BRATS’2017 dataset [50, 51]). (a) T2-FLAIR MRI, (b) manual “ground truth” glioma segmentation by an experienced board-certified radiation oncologist, (c) machine learning—SVM model glioma segmentation [52], and (d) both, manual and ML, segmented annotations overlap; for four different subjects.
Overall, tumor/target segmentation/auto-contouring using ML-based methods still remains challenging for some reasons such as availability of big data of multimodal images with their “ground truth” annotation data for training these models. Recent advances in computer vision, specifically around deep learning [61], are particularly well suited for segmentation and it has shown superiority over the other machine learning algorithms for tumor and organs segmentation tasks.
\nThe planning process starts by delineating both the target(s) and the OARs as we discussed it earlier in the image segmentation section (Section 2.2.4). Once the target volumes and OARs have been outlined/contoured, the planning process continues by (1) setting dosimetric goals for targets and normal tissues; (2) selecting an appropriate treatment technique (e.g., 3D, fixed beam IMRT, VMAT (volumetric arc radiation therapy), protons); (3) iteratively modifying the beams/weights/etc., until the planning goals have been achieved; and (4) evaluating (estimating the treatment dose distributions with prescribed doses in the treatment planning system using dose calculation algorithms) and approving the plan [14]. The applications of machine learning in radiotherapy treatment planning as a tool for knowledge-based treatment planning (KBTP) and automated/self-driven planning process will be discussed in this section.
\nPrior information about patient status and previously archived treatment plans, particularly if performed by expert medical dosimetrists/physicists, could be used to inform the treating team of a currently pending case [2]. This concept of using prior treatment planning information constitutes the underlying principle of the so-called knowledge-based treatment planning. Such KBTP approaches have leveraged hundreds of prior treatment plans to reproducibly improve planning efficiency across multiple disease sites [62]. Figure 8 illustrates the schematic of a KBTP System [2]. The motivation for KBTP approach lies in reducing current complexity and time spent on generating a new treatment plan from each incoming patient, as well as its potential for decision-making support in radiotherapy.
\nSchematic of a KBTP system. Initially, the user builds a query using features related to patient, disease, imaging, treatment setup, dose, etc., for the treatment plan (TP). Then, the database returns a set of similar treatment plans that the user could select from to optimize and compare with the current one according to the query (reproduced from [2]).
Several studies [63, 64, 65, 66, 67] have carried out to explore the utilization of KBTP approach for treatment plan generation in radiotherapy. The current scientific research and available commercial products for KBTP are limited to predicting DVHs within accepted ranges [14]. Plans generated based on KBTP utilizing artificial intelligence often meet or exceed adherence to dose constraints compared to manually generated plans in many clinical scenarios (e.g., prostate cancer [63], cervical cancer [64], gliomas and meningiomas [65], head and neck cancer [66], and spine SBRT [67]). A more recent commercial product, Quick Match (Siris Medical, Redwood City, CA, USA), uses gradient boosting (the most accurate algorithm on expectation when structured data are available) to explore predictions in dosimetric trade-offs [68]. This application provides quick rough predicted treatment planning results to be obtained before the treatment planning process. Thus it can facilitate communication between dosimetrist and physicians, establish individualized and achievable goals, and help physicians and patients decide the course of a plan before initializing the treatment planning process. For example, it can help to choose an optimal technique (e.g., photon versus protons). This approach has also been applied to post-planning quality assurance of DVH data [69, 70].
\nOverall, the incentive for such an approach like KBTP lies in reducing current complexity and time spent on generating a new treatment plan from each incoming patient. It is believed that such a standardization process based on KBTP can help enhance consistency, efficiency, and plan quality. Ultimately, data-driven planning is not fully automated at present as it requires expert oversight and/or intervention to ensure safely deliverable treatment plans.
\nOnce the dosimetric goals have been established and the technique chosen, automatic plan generation is also possible [14].
\nSome attempts [71, 72] have made to solve various aspects of this problem by predicting the best beam orientations. The larger task of automated treatment planning, however, is well suited for reinforcement learning method [14]. Reinforcement is extensively used in games, self-driving cars, and other popular-culture applications. In reinforcement learning method, an algorithm learns to navigate a set of rules, given some constraints, by self-correcting its decisions. Basically, the algorithm will take a decision (for instance, increase the weight of a given constraint) and learn from the simulator (the treatment planning system) whether the decision resulted in the right direction [14]. This technique has successfully used by Google Brain to develop an algorithm capable of beating a Go world champion [73]. So, reinforcement technique could provide performance at the level of our best dosimetrists if properly implemented.
\nOverall, one challenge of achieving full automatic planning using reinforcement learning lies in the close integration and need for robust treatment planning systems (TPSs) [14]. The future vision is toward a fully-automated planning process, from contouring to plan creation [62], with the human experts (dosimetrists, physicists, and physicians) evaluating, supervising, and providing QA to the given results.
\nQuality assurance (QA) is demanding for the safe delivery of radiotherapy. It represents a core part of a medical physicist’s task in the clinical practice. Machine learning could be utilized to solve multiple long-standing problems and improve workflow efficiency. Its applications in the quality assurance (e.g., detection and prediction of radiotherapy errors, and treatment planning QA) and treatment delivery validation (e.g., prediction planning deviations from the initial intentions, and prediction the need for re-planning for adaptive radiotherapy) are discussed in this section.
\nMachine learning has potential in many aspects of radiotherapy QA program, specifically in error detection and prevention, treatment machine QA, patient-specific quality assurance, etc. In addition, ML may contribute to automating the QA process and analysis, which significantly influence an increase in efficiency and a decrease in the physical effort in performing the QA.
\nNumerous studies [74, 75, 76, 77, 79, 80, 81, 82, 83] have conducted to develop a computerized system for QA process based on machine learning methods. We can generally categorize these QA into the machine-based and patient-based approach. For machine-based QA approach, ML utilizations for automatic QA process of medical linear accelerator (Linac) machine [74, 75, 76, 77] have investigated by research scientists. A study by Li et al. [74] investigated the application of ANN to monitor the performance of the Linac for continuous improvement of patient safety and quality of care. The preliminary results showed better accuracy and effective applicability in the dosimetry and QA field over other techniques, and in some cases, its performance beat the detection rate by current clinical metrics. El Naqa et al. [75] introduced a system utilizing anomaly detection to overcome the problem of direct modeling of QA errors and rare events in radiotherapy and to support the intent of automated QA and safety management for patients undergo radiotherapy treatment. Ford et al. [76] and Hoisak et al. [77] investigated quantifying the error-detection effectiveness of commonly used quality control (QC) measures [76] preventative maintenance [77] in radiation oncology. The results indicated that the effectiveness of QC measures in radiation oncology depends sensitively on which checks are used and in which combinations [76], and also a decreased machine downtime and other technical failures leading to treatment cancellations [77]. The ability of these ML algorithms to automatically detect outliers allows physicists to focus attention on those aspects of a process most likely to impact the patient care, as recommended in AAPM Task Group report 100 [78].
\nFor patient-based QA approach, application of ML algorithms for a plan and patient-specific QA, multi-leaf collimators (MLCs) QA, and imaging [79, 80, 81, 82, 83] have discovered by many investigators. A study by Valdes et al. [80] investigated the use of SVM-based system to automatically detect problems with the Linac 2D/3D imaging system that are used for patient IGRT treatment accuracy. The proposed method results showed that the bare minimum and the best practice QA programs could be implemented with the same manpower. Regarding plan QA and patient-specific QA, investigators [81, 82] studied applications of Poisson regression with LASSO regularization to predict individualized IMRT QA passing rates. Their results pointed out that virtual IMRT QA can predict passing rates with a high likelihood, allows the detection of failures due to setup errors. Osman et al. [79] and Carlson et al. [83] utilized NN and a cubist algorithm, respectively, to predict MLC positional errors using the Linac generated log file data of IMRT and VMAT delivered plans. Their studies results showed that predicted parameters were in closer agreement to the delivered parameters than the planned parameters. The inclusion of these predicted deviations in leaves positioning into the TPS during dose calculation leads to a more realistic representation of plan delivery. Figure 9 illustrates a generic flow diagram and results of an NN utilized for prediction of MLCs positional errors [79].
\nTop: A generic flow diagram of the proposed method of prediction MLC positional errors [79]. Bottom: Differences in the leaf positions between the delivered and planned (upper), and delivered and predicted with NN (lower). Boxes report quartiles including the median (the 50% central sample distribution); whiskers and dots indicate outliers.
Overall, despite these significant improvements in QA processes with the involvement of ML, they carry implicit maintenance costs in the form of additional QA demands for the algorithms themselves. The performance of all deployed ML-based algorithms will, therefore, need to be verified periodically using an evolving series of tests [62]. Virtual QA can have profound implications on the current IMRT/VMAT process and potentially enabling intelligent resource allocation in favor of plans more likely to fail.
\nTumor shrinkage and anatomical patient variations (e.g., due to weight loss) may occur throughout a few weeks of a fractionated radiotherapy treatment. Adaptive radiation therapy (ART) is a treatment approach that uses frequent imaging to compensate for anatomical differences that occur during the course of treatment. Images are taken daily, or almost daily. When significant changes are observed, replanning is considered. It is possible to achieve image-guided adaptation either off-line (i.e., using image information acquired during a fraction for improving following fraction) or online (i.e., changing treatment plan for a fraction based on information from the same fraction).
\nThe re-planning process involves three steps [84]: (1) simulating the plan from the daily CBCT image dataset to calculate the estimated actual delivered daily dose for the given treatment fraction, (2) delineating the structures of interest to obtain daily DVHs to provide dose metrics for the tumor and OARs from which radiation oncologists can evaluate treatment plan effectiveness, and (3) modifying the doses to the therapeutic target and OARs to meet the dose constraints in the original treatment plan. The implementation of adaptive radiotherapy into routine clinical practice is technically challenging and requires significant resources to perform and validate each process step. It needs to be fast (where time is a big issue) in order to fit into the clinical workflow. Machine learning techniques, i.e., deep learning, may offer potentials to have very sophisticated software tools for adaptive therapy. In recent years, deep learning [61] applications have grown in a variety of fields including video games, computer vision, and pattern recognition.
\nA number of researchers [85, 86, 87, 88] have investigated the application of ML, particularly deep learning, in treatment re-planning process for adaptive radiotherapy. Studies by Guidi et al. [85] and Chetvertkov et al. [86] conducted to predict patients who would benefit from ART and re-planning intervention using SVM [85] and PCA [86] ML models. The studies results indicated a capability of identifying patients would benefit from ART and ideal time for a re-planning intervention. Tseng et al. [87] investigated deep reinforcement learning based on historical treatment plans for developing automated radiation adaptation protocols for lung cancer patients aiming to maximize tumor local control at reduced rates of radiation pneumonitis. The study findings revealed that automated dose adaptation by deep reinforcement learning is a feasible and promising approach for achieving similar results to those chosen by clinicians. Varfalvy et al. [88] introduced a new automated patient classification method based on relative gamma analysis and hidden Markov models to identify patients undergoing important anatomical changes during radiotherapy. The results obtained indicated that it can complement the clinical information collected during treatment and help identify patients in need of a plan adaptation.
\nOverall, adaptive radiotherapy demands a high-speed planning system, combined with high-quality imaging. Deep learning-based ML methods have shown potential and feasibility to transform adaptive radiation therapy more effectively and efficiently into the routine clinical practice soon. Effective implementation of adaptive radiation therapy can further improve the precision in the radiotherapy treatments.
\nPatient follow-up begins at the start of the treatment and continues to beyond the end of the treatment. Accurate prediction of treatment outcomes would provide clinicians with better tools for informed decision-making about expected benefits versus anticipated risks [2]. Machine learning has the potential to revolutionize the way radiation oncologists follow patients treated with definitive radiation therapy [14]. In addition, it may potentially enable practical use of precision medicine in radiation oncology by predicting treatment outcomes for individual patients using radiomics “tumor/healthy tissue phenotypes” analysis.
\nRadiotherapy treatment outcomes are determined by complex interactions among treatment, anatomical, and patient-related variables [2]. A key component of radiation oncology research is to predict at the time of treatment planning, or during the course of fractionated radiation treatment, the tumor control probability (TCP) and normal tissue control probability (NTCP) for the type of treatment being considered for that particular patient [2]. Recent approaches have utilized increasingly data-driven models incorporating advanced bioinformatics and machine learning tools in which dose-volume metrics are mixed with other patients- or disease-based prognostic factors in order to improve outcomes prediction [2]. Obviously, better models based on early assessment are needed to predict the outcome, in time for treatment intensification with additional radiotherapy, early addition of systemic therapy, or application of a different treatment modality [14].
\nMany research scientists [89, 90, 91, 92, 93, 94, 95] have investigated the application of ML in radiotherapy treatment response and outcome predictions. Lee et al. [89] studied utilizing of Bayesian network ensemble to predict radiation pneumonitis risk for NSCLC patients whom received curative 3D conformal radiotherapy. The preliminary results demonstrated that such framework combined with an ensemble method can possibly improve the prediction of radiation pneumonitis under real-life clinical circumstances. Naqa et al. [90] introduced a data mining framework estimating model parameters for predicting TCP using statistical resampling and a logistic, SVM, logistic regression, Poisson-based TCP, and cell kill equivalent uniform dose model. Their findings indicated that prediction of treatment response can be improved by utilizing data mining approaches, which were able to unravel important non-linear complex interactions among model variables and have the capacity to predict on unseen data for prospective clinical applications. Zhen et al. [91] introduced a CNN model to analyze the rectum dose distribution and predict rectum toxicity. The evaluation results demonstrated the feasibility of building a CNN-based rectum dose-toxicity prediction model with transfer learning for cervical cancer radiotherapy. Deist et al. [92] studied the comparison of six ML classifiers (namely, decision tree, random forest, NN, SVM, elastic net logistic regression, and LogitBoost) for chemo-radiotherapy to estimate their average discriminative performance for radiation treatment outcome prediction. The study results indicated that random forest and elastic net logistic regression yield higher discriminative performance in (chemo) radiotherapy outcome and toxicity prediction than other studied classifiers. Yahya et al. [93] explored multiple statistical-learning strategies for prediction of urinary symptoms following external beam radiotherapy of the prostate. The study results showed that logistic regression and multivariate adaptive regression splines (MARS) were most likely to be the best-performing strategy for the prediction of urinary symptoms. Zhang et al. [94] studied the prediction of organ-at-risk complications as a function of dose-volume constraint settings using SVMs and decisions trees. Their results showed that ML can be used for predicting OAR complications during treatment planning allowing for alternative dose-volume constraint settings to be assessed within the IMRT planning framework. A review by Kang et al. [95] presented the use of ML to predict radiation therapy outcomes from the clinician’s point of view. The study focused on three popular ML methods: logistic regression, SVM, and ANN. The study concluded that although current studies are in exploratory stages, the overall methodology has progressively matured, and the field is ready for larger-scale further investigation.
\nOverall, a significant hope of advanced clinical informatics systems would be the potential to learn even more about the safety and effectiveness of the therapies that are provided to patients. The rapid adoption of technological advancements in radiotherapy has made outcomes analyses of both treatment regimens and the systems that deliver them to be separated substantially in time. Successful application of advanced ML tools for radiation oncology big data is essential to better-predicting radiotherapy treatment response and outcomes. The ultimate measure of success is an improvement in outcomes which can manifest as decreased toxicity or increased tumor control.
\nPrecision medicine is a treatment strategy for making decisions about a molecularly targeted agent according to genetic mutations, rather than affected organs. Radiomics is the comprehensive quantitative analysis of medical images in order to extract a large number of phenotypic features (including those based on size and shape, image intensity, texture, relationships between voxels, and fractal characteristics) reflecting cancer traits or phenotypes. Then it explores the associations between the features and patients’ prognoses in order to improve decision-making at each radiation treatment step (diagnosis, treatment planning, treatment delivery, and follow-up) and hence precision medicine in radiotherapy [96]. Individual patients can be stratified into subtypes based on radiomic biomarkers that contain information about cancer traits that determine the patient’s prognosis [97]. Machine-learning algorithms can then be deployed to correlate the computer-extracted image-based features in radiomics with biological observations or clinical outcomes. Here, we present some current results and emerging paradigms in radiomics boosted with ML approaches in clinical radiation oncology (recently received higher attention from the investigators) to maximize its potential impact on precision radiotherapy.
\nSeveral research scientists [97, 98, 99, 100, 101, 102] have investigated the using of ML methods for predicting radiotherapy outcomes (e.g., survival, treatment failure or recurrence, toxicity or developed a late complication, etc.) using radiomics features to improve decision-making for precision medicine. A review study by Arimura et al. [97] showed that radiomic approaches in combination with AI may potentially enable the practical use of precision medicine in radiation therapy by predicting outcomes and toxicity for individual patients. Aerts et al. [98] performed a radiomic analysis of 440 features quantifying tumor image intensity, shape, and texture, which are extracted from CT data of patients with lung or head-and-neck cancer. The study findings proved the power of radiomics for identifying a general prognostic phenotype existing in both lung and head-and-neck cancer. Figure 10 shows a workflow of radiomics analysis (example: CT radiomic analysis of with lung cancer) [98]. A study by Depeursinge et al. [99] investigated the importance of pre-surgical CT intensity and texture information from ground-glass opacities and solid nodule components for the prediction of adenocarcinoma recurrence in the lung using LASSO and SVMs, and their survival counterparts: Cox-LASSO and survival SVMs. The study results showed the usefulness of the method in clinical practice to identify patients for which no recurrence is expected with very high confidence using a pre-surgical CT scan only. Lambin et al. [100] studied the development of automated and reproducible analysis methodologies to extract more information from image-based features. The study addressed the radiomics as one of the approaches that hold great promises but need further validation in multi-centric settings. A review by Wu et al. [101] recommended that ultimately prospective validation in multi-center clinical trials will be needed to demonstrate the clinical validity and utility of newly identified imaging markers and truly establish the value of radiomics and radiogenomics in precision radiotherapy. Lao et al. [102] investigated if deep features extracted via transfer learning can generate radiomics signatures for prediction of overall survival in patients with glioblastoma multiforme using the LASSO Cox regression model. The study outcomes demonstrated that the proposed method is capable to generate prognostic imaging signature for OS prediction and patient stratification for glioblastoma, indicating the potential of deep imaging feature-based biomarker in preoperative care of glioblastoma patients.
\nA typical radiomics workflow. Imaging: Tumors are different. Example CT images with tumor contours of lung cancer patients. Segmentation: 3D visualizations of tumor contours delineated by experienced physicians on all CT slices. Pre-processing: Strategy for extracting radiomics data from images. Feature Extraction: Features are extracted from the defined tumor contours on the CT images quantifying tumor intensity, shape, texture and wavelet texture. Analysis: For the analysis, the radiomics features are compared with clinical data and gene-expression data (reproduced from [98]).
Overall, radiomics is the study of imaging data from any imaging source that is used to predict the therapeutic outcome, as well as radiogenomics. The limited reproducibility of imaging systems both within and across institutions remains a significant challenge for radiomics [98, 100]. Application of deep learning to image quantification has produced stellar results in other areas [103] which can be transferred into the radiomics analysis. Physicians may prescribe a more or less intense radiation regimen for an individual based on model predictions of local control benefit and toxicity risk [2], which would be considered for the optimal treatment planning design process and hence improving the quality of life for radiotherapy cancer patients. Also, as imaging is routinely used in clinical practice, radiomics is providing an unprecedented opportunity to improve decision-making support toward precision medicine in cancer treatment at low cost.
\nA comprehensive review of the most recent evolution and ongoing research utilizing machine learning methods in radiation oncology in the era of big data for precision medicine has been provided in this chapter and critically discussed.
\nThere are ongoing community-wide efforts in term of big data in radiation oncology, e.g., [9, 10, 50, 51] have made available and established validation frameworks [50] used as a benchmark for the evaluation of different algorithms. Deep learning [61] based models have indicated superiority among the other alternatives for the most prediction tasks in radiation oncology. However, it requires a lot of annotated datasets (across multiple institutions) to tune the algorithm (even when transfer learning is used [14]) to obtain high prediction accuracy. This can prove challenging in radiation oncology, where datasets are limited. Standardizing the radiation oncology nomenclature (i.e., clinical, dosimetric, imaging, etc.), which is aided by the AAPM task group TG-263 efforts [104], and developing standards for data collection process (structures) of the patient data are also essential for training models using datasets from multiple institutions.
\nThere is no one algorithm works best for every problem (“No Free Lunch”). Each ML algorithm has its strengths and limitations. Table 1 lists the strengths and weaknesses of the most machine learning methods discussed here appearing in radiation oncology studies. It is believed that such usage optimization of these models with available resources would provide improved solutions. A major limitation in the acceptance of ML by the larger medical community has been addressed as the “black box” stigma, where the ML algorithm maps a given input data to output predictions without providing any additional insight into the system mapping [6]. Interpretability of algorithms used (e.g., the ability for humans experts to understand the reasons behind a prediction) will play an important role to avoid preventable errors. Although there are inherently interpretable ML algorithms, for instance, decision trees, Bayesian networks, or generalized linear models (e.g., logistic regression), they are usually outperformed in terms of accuracy by ensemble methods or deep neural networks (not interpretable and provide very little insight) for large datasets [6, 13]. The development of accurate and interpretable models using different ML architectures is an active area of research [6]. As with any algorithm that we use in radiation oncology today (e.g., dose calculation or deformable registration), ML algorithms will need acceptance, commissioning, and QA to ensure that the right algorithm or model are applied to the right application and that the model results make sense in a given clinical situation. Finally, the field of radiation oncology is highly algorithmic and data-centric, and while the road ahead is filled with potholes, the destination holds tremendous promise [14].
\nMethod | \nStrengths | \nWeaknesses | \n
---|---|---|
Decision tree | \nInterpretability (with a format consistent with many clinical pathways) | \nOvergrowing a tree with too few observations at leaf nodes | \n
Random forest | \nOften can produce very accurate predictions with little feature engineering | \nNot easily interpretable, and not optimizing the number of trees | \n
LASSO regression | \nBetter interpretability (compared to ridge regularization method) | \nProvides a bias towards zero (not be appropriate in some applications) | \n
Gradient boosting machines | \nGenerates very stable results (compared to random forest) | \nMore tuning parameters (compared to random forest), and overfitting | \n
Support vector machines | \nVery accurate, few parameters that require tuning, and kernels options | \nNot readily interpretable, and not optimizing the parameters perfectly | \n
Neural networks or more precisely artificial neural networks | \nWorks even if one or a few units fail to respond to the network | \nReferred to as “black box” models and provide very little insight, and require a large diversity of training datasets | \n
Deep learning | \nVery accurate, can be adapted to many types of problems, and the hidden layers reduce the need for feature engineering | \nRequires a very large amount of data, and computationally intensive to train | \n
Logistic regression | \nHave a nice probabilistic interpretation, and updated easily with new data | \nNot flexible enough to naturally capture more complex relationships | \n
K-means | \nFast, simple, and flexible | \nManually specify the number of clusters | \n
Ensembles (decision tree) | \nPerform very well, robust to outliers, and scalable | \nUnconstrained, and prone to overfitting | \n
Principal component analysis | \nVersatile, fast, and simple to implement | \nNot interpretable, and manually set a threshold for a cumulative variance | \n
Naive Bayes | \nPerforms surprisingly well, easy to implement, and can scale with the dataset | \nOften beaten by models properly trained and tuned (algorithms listed) | \n
Strengths and weaknesses of the most machine learning methods discussed here appearing in radiation oncology studies.
The reported prediction results [15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 41, 42, 43, 44, 45, 46, 47, 52, 53, 54, 55, 56, 57, 58, 59, 60, 63, 64, 65, 66, 67, 71, 72, 74, 75, 76, 77, 79, 80, 81, 82, 83, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 97, 98, 99, 100, 101, 102] by investigators indicate the performance of these predictive models on data that used in modeling. However, these ML models can suffer from different data biases which may lead to lack of generalizability. A machine learning system trained on local datasets only may not be able to predict (reproduce) the needs of out-of-sample datasets (new datasets that are not presented in the training data). External validation of models in cohorts, which were acquired independently from the discovery cohort (e.g., from another institution) is considered the gold standard for true estimates of performance and generalizability of prediction models [6]. The application of different algorithms to the same dataset may yield variable results for predictors found to be significantly associated with the outcome of interest [6, 105]. However, this may also suggest a potential limitation of self-critical assessment of published ML models or realistic confidence levels with implications for their practical clinical value [6].
\nAlthough promising and improving accuracy results of many ML-based predictive models in radiation oncology have been reported [18, 19, 21, 31, 32, 33, 34, 35, 36, 37, 38, 41, 42, 43, 53, 54, 55, 74, 79, 80, 81, 82, 83, 85, 86, 89, 90, 91, 92, 93, 94, 95, 97, 98, 99, 100, 101, 102], the effective applications of these methods in day-to-day clinical practice are very few yet. Such an example of a recently deployed commercial product into clinical use is Quick Match (Siris Medical, Redwood City, CA, USA) [68]. A private initiative, such as IBM’s Watson, is already used in some institutions such as the Memorial Sloan Kettering Cancer Center in New York [106, 107, 108, 109]. Watson Oncology [108] is a cognitive AI computing system designed to support the broader oncology community of physicians as they consider treatment options with their patients. To improve the prediction accuracy of these reported results, more training and validation datasets from multi-institution are required. Such frameworks, e.g., [50] to compare these methods on standard consensus data to establish benchmarks for evaluating different models would definitely lead to improving these results and developing robust toolkits/systems. It is anticipated to see ML and AI tools very soon settled more effectively with the indispensable role in the routine clinical practice for the benefit of patients, society, and the profession.
\nThe machine learning systems have been developed and deployed to do jobs on their own. Automated clinical processes in radiation oncology could be auto-piloted with driving technologies to execute automated tasks. For example, data-driven planning [63, 64, 65, 66, 67] is not fully automated at present as it requires expert oversight and/or intervention to ensure safely deliverable treatment plans. One challenge of achieving full automatic planning using reinforcement learning lies in the close integration and need for robust TPSs [14]. The future vision is toward a fully-automated planning process, from contouring to plan creation. Machine-based and patient-based virtual QA can have profound implications on the current IMRT/VMAT process. The automated process nature would definitely lead to expediting radiation oncology workflow and reduce the time burden of human intervention [62].
\nML tools for computer-aided detection/diagnosis [15, 16, 17, 18, 19, 20, 21, 22] as “second opinion” systems for clinical decision-making support would undoubtedly enhance the radiologists’ performance and hence improved diagnostic performance. The emerging paradigms in radiomics for therapeutic outcome predictions (i.e., patient’s survival, decrease recurrence, late complication, etc.) [97, 98, 99, 100, 101, 102] for individual patients would maximize its potential impact on precision radiotherapy. Individual patients can be stratified into subtypes based on radiomic biomarkers that contain information about cancer traits that determine the patient’s prognosis [97]. Therefore, physicians may prescribe a more or less intense radiation regimen for an individual based on model predictions of local control benefit and toxicity risk [2], which would be considered for the optimal treatment planning design process and hence improving the quality of life for radiotherapy cancer patients. Effective implementation of adaptive radiation therapy with ML [85, 86, 87, 88] can also further improve the precision in the radiotherapy treatments. The pre-planning prediction of dosimetric tradeoffs to assist physicians and patients to make better informed decisions about treatment modality and dose prescription [68] thus it can establish individualized and achievable goals. The clinical implications derived from personalized cancer therapy ensure not only that patients receive optimal treatment, but also that the right resources are being used for the right patients.
\nMachine learning methods used in radiation oncology workflow, from patient consult to follow-up, are presented and discussed in this chapter. Big data in radiation oncology, efforts made and current challenges, are addressed. With the era of big data, the utilization of machine learning algorithms in radiation oncology is growing fast. ML techniques could compensate for human limitations in handling a large amount of flowing information in an efficient manner, in which simple errors can make the difference between life and death. Machine learning is also indispensable in the radiomics scheme, characterization of image phenotypes of the tumor, with the potential for decision-making and precision medicine in radiation therapy by predicting treatment outcomes for individual patients rather than one-size-fits-all approach.
\nThe author is grateful to the attending physicians, physicists, residents, and staff at the radiation oncology department, at American University of Beirut Medical Center (AUBMC), Lebanon. Most of the clinical aspects provided in this chapter were based on the author’s knowledge and experience gained during his residency at AUBMC. The contents are solely representing the author’s view. The author also specially thanks the IntechOpen for granting this chapter a full funding for Open-Access publication.
\nThe author has no conflict of interest.
Edited by Jan Oxholm Gordeladze, ISBN 978-953-51-3020-8, Print ISBN 978-953-51-3019-2, 336 pages,
\nPublisher: IntechOpen
\nChapters published March 22, 2017 under CC BY 3.0 license
\nDOI: 10.5772/61430
\nEdited Volume
This book serves as a comprehensive survey of the impact of vitamin K2 on cellular functions and organ systems, indicating that vitamin K2 plays an important role in the differentiation/preservation of various cell phenotypes and as a stimulator and/or mediator of interorgan cross talk. Vitamin K2 binds to the transcription factor SXR/PXR, thus acting like a hormone (very much in the same manner as vitamin A and vitamin D). Therefore, vitamin K2 affects a multitude of organ systems, and it is reckoned to be one positive factor in bringing about "longevity" to the human body, e.g., supporting the functions/health of different organ systems, as well as correcting the functioning or even "curing" ailments striking several organs in our body.
\\n\\nChapter 1 Introductory Chapter: Vitamin K2 by Jan Oxholm Gordeladze
\\n\\nChapter 2 Vitamin K, SXR, and GGCX by Kotaro Azuma and Satoshi Inoue
\\n\\nChapter 3 Vitamin K2 Rich Food Products by Muhammad Yasin, Masood Sadiq Butt and Aurang Zeb
\\n\\nChapter 4 Menaquinones, Bacteria, and Foods: Vitamin K2 in the Diet by Barbara Walther and Magali Chollet
\\n\\nChapter 5 The Impact of Vitamin K2 on Energy Metabolism by Mona Møller, Serena Tonstad, Tone Bathen and Jan Oxholm Gordeladze
\\n\\nChapter 6 Vitamin K2 and Bone Health by Niels Erik Frandsen and Jan Oxholm Gordeladze
\\n\\nChapter 7 Vitamin K2 and its Impact on Tooth Epigenetics by Jan Oxholm Gordeladze, Maria A. Landin, Gaute Floer Johnsen, Håvard Jostein Haugen and Harald Osmundsen
\\n\\nChapter 8 Anti-Inflammatory Actions of Vitamin K by Stephen J. Hodges, Andrew A. Pitsillides, Lars M. Ytrebø and Robin Soper
\\n\\nChapter 9 Vitamin K2: Implications for Cardiovascular Health in the Context of Plant-Based Diets, with Applications for Prostate Health by Michael S. Donaldson
\\n\\nChapter 11 Vitamin K2 Facilitating Inter-Organ Cross-Talk by Jan O. Gordeladze, Håvard J. Haugen, Gaute Floer Johnsen and Mona Møller
\\n\\nChapter 13 Medicinal Chemistry of Vitamin K Derivatives and Metabolites by Shinya Fujii and Hiroyuki Kagechika
\\n"}]'},components:[{type:"htmlEditorComponent",content:'This book serves as a comprehensive survey of the impact of vitamin K2 on cellular functions and organ systems, indicating that vitamin K2 plays an important role in the differentiation/preservation of various cell phenotypes and as a stimulator and/or mediator of interorgan cross talk. Vitamin K2 binds to the transcription factor SXR/PXR, thus acting like a hormone (very much in the same manner as vitamin A and vitamin D). Therefore, vitamin K2 affects a multitude of organ systems, and it is reckoned to be one positive factor in bringing about "longevity" to the human body, e.g., supporting the functions/health of different organ systems, as well as correcting the functioning or even "curing" ailments striking several organs in our body.
\n\nChapter 1 Introductory Chapter: Vitamin K2 by Jan Oxholm Gordeladze
\n\nChapter 2 Vitamin K, SXR, and GGCX by Kotaro Azuma and Satoshi Inoue
\n\nChapter 3 Vitamin K2 Rich Food Products by Muhammad Yasin, Masood Sadiq Butt and Aurang Zeb
\n\nChapter 4 Menaquinones, Bacteria, and Foods: Vitamin K2 in the Diet by Barbara Walther and Magali Chollet
\n\nChapter 5 The Impact of Vitamin K2 on Energy Metabolism by Mona Møller, Serena Tonstad, Tone Bathen and Jan Oxholm Gordeladze
\n\nChapter 6 Vitamin K2 and Bone Health by Niels Erik Frandsen and Jan Oxholm Gordeladze
\n\nChapter 7 Vitamin K2 and its Impact on Tooth Epigenetics by Jan Oxholm Gordeladze, Maria A. Landin, Gaute Floer Johnsen, Håvard Jostein Haugen and Harald Osmundsen
\n\nChapter 8 Anti-Inflammatory Actions of Vitamin K by Stephen J. Hodges, Andrew A. Pitsillides, Lars M. Ytrebø and Robin Soper
\n\nChapter 9 Vitamin K2: Implications for Cardiovascular Health in the Context of Plant-Based Diets, with Applications for Prostate Health by Michael S. Donaldson
\n\nChapter 11 Vitamin K2 Facilitating Inter-Organ Cross-Talk by Jan O. Gordeladze, Håvard J. Haugen, Gaute Floer Johnsen and Mona Møller
\n\nChapter 13 Medicinal Chemistry of Vitamin K Derivatives and Metabolites by Shinya Fujii and Hiroyuki Kagechika
\n'}]},successStories:{items:[]},authorsAndEditors:{filterParams:{sort:"featured,name"},profiles:[{id:"6700",title:"Dr.",name:"Abbass A.",middleName:null,surname:"Hashim",slug:"abbass-a.-hashim",fullName:"Abbass A. Hashim",position:null,profilePictureURL:"https://mts.intechopen.com/storage/users/6700/images/1864_n.jpg",biography:"Currently I am carrying out research in several areas of interest, mainly covering work on chemical and bio-sensors, semiconductor thin film device fabrication and characterisation.\nAt the moment I have very strong interest in radiation environmental pollution and bacteriology treatment. The teams of researchers are working very hard to bring novel results in this field. I am also a member of the team in charge for the supervision of Ph.D. students in the fields of development of silicon based planar waveguide sensor devices, study of inelastic electron tunnelling in planar tunnelling nanostructures for sensing applications and development of organotellurium(IV) compounds for semiconductor applications. I am a specialist in data analysis techniques and nanosurface structure. I have served as the editor for many books, been a member of the editorial board in science journals, have published many papers and hold many patents.",institutionString:null,institution:{name:"Sheffield Hallam University",country:{name:"United Kingdom"}}},{id:"54525",title:"Prof.",name:"Abdul Latif",middleName:null,surname:"Ahmad",slug:"abdul-latif-ahmad",fullName:"Abdul Latif Ahmad",position:null,profilePictureURL:"//cdnintech.com/web/frontend/www/assets/author.svg",biography:null,institutionString:null,institution:null},{id:"20567",title:"Prof.",name:"Ado",middleName:null,surname:"Jorio",slug:"ado-jorio",fullName:"Ado Jorio",position:null,profilePictureURL:"//cdnintech.com/web/frontend/www/assets/author.svg",biography:null,institutionString:null,institution:{name:"Universidade Federal de Minas Gerais",country:{name:"Brazil"}}},{id:"47940",title:"Dr.",name:"Alberto",middleName:null,surname:"Mantovani",slug:"alberto-mantovani",fullName:"Alberto Mantovani",position:null,profilePictureURL:"//cdnintech.com/web/frontend/www/assets/author.svg",biography:null,institutionString:null,institution:null},{id:"12392",title:"Mr.",name:"Alex",middleName:null,surname:"Lazinica",slug:"alex-lazinica",fullName:"Alex Lazinica",position:null,profilePictureURL:"https://mts.intechopen.com/storage/users/12392/images/7282_n.png",biography:"Alex Lazinica is the founder and CEO of IntechOpen. After obtaining a Master's degree in Mechanical Engineering, he continued his PhD studies in Robotics at the Vienna University of Technology. Here he worked as a robotic researcher with the university's Intelligent Manufacturing Systems Group as well as a guest researcher at various European universities, including the Swiss Federal Institute of Technology Lausanne (EPFL). During this time he published more than 20 scientific papers, gave presentations, served as a reviewer for major robotic journals and conferences and most importantly he co-founded and built the International Journal of Advanced Robotic Systems- world's first Open Access journal in the field of robotics. Starting this journal was a pivotal point in his career, since it was a pathway to founding IntechOpen - Open Access publisher focused on addressing academic researchers needs. Alex is a personification of IntechOpen key values being trusted, open and entrepreneurial. Today his focus is on defining the growth and development strategy for the company.",institutionString:null,institution:{name:"TU Wien",country:{name:"Austria"}}},{id:"19816",title:"Prof.",name:"Alexander",middleName:null,surname:"Kokorin",slug:"alexander-kokorin",fullName:"Alexander Kokorin",position:null,profilePictureURL:"https://mts.intechopen.com/storage/users/19816/images/1607_n.jpg",biography:"Alexander I. Kokorin: born: 1947, Moscow; DSc., PhD; Principal Research Fellow (Research Professor) of Department of Kinetics and Catalysis, N. Semenov Institute of Chemical Physics, Russian Academy of Sciences, Moscow.\r\nArea of research interests: physical chemistry of complex-organized molecular and nanosized systems, including polymer-metal complexes; the surface of doped oxide semiconductors. He is an expert in structural, absorptive, catalytic and photocatalytic properties, in structural organization and dynamic features of ionic liquids, in magnetic interactions between paramagnetic centers. The author or co-author of 3 books, over 200 articles and reviews in scientific journals and books. He is an actual member of the International EPR/ESR Society, European Society on Quantum Solar Energy Conversion, Moscow House of Scientists, of the Board of Moscow Physical Society.",institutionString:null,institution:{name:"Semenov Institute of Chemical Physics",country:{name:"Russia"}}},{id:"62389",title:"PhD.",name:"Ali Demir",middleName:null,surname:"Sezer",slug:"ali-demir-sezer",fullName:"Ali Demir Sezer",position:null,profilePictureURL:"https://mts.intechopen.com/storage/users/62389/images/3413_n.jpg",biography:"Dr. Ali Demir Sezer has a Ph.D. from Pharmaceutical Biotechnology at the Faculty of Pharmacy, University of Marmara (Turkey). He is the member of many Pharmaceutical Associations and acts as a reviewer of scientific journals and European projects under different research areas such as: drug delivery systems, nanotechnology and pharmaceutical biotechnology. Dr. Sezer is the author of many scientific publications in peer-reviewed journals and poster communications. Focus of his research activity is drug delivery, physico-chemical characterization and biological evaluation of biopolymers micro and nanoparticles as modified drug delivery system, and colloidal drug carriers (liposomes, nanoparticles etc.).",institutionString:null,institution:{name:"Marmara University",country:{name:"Turkey"}}},{id:"61051",title:"Prof.",name:"Andrea",middleName:null,surname:"Natale",slug:"andrea-natale",fullName:"Andrea Natale",position:null,profilePictureURL:"//cdnintech.com/web/frontend/www/assets/author.svg",biography:null,institutionString:null,institution:null},{id:"100762",title:"Prof.",name:"Andrea",middleName:null,surname:"Natale",slug:"andrea-natale",fullName:"Andrea Natale",position:null,profilePictureURL:"//cdnintech.com/web/frontend/www/assets/author.svg",biography:null,institutionString:null,institution:{name:"St David's Medical Center",country:{name:"United States of America"}}},{id:"107416",title:"Dr.",name:"Andrea",middleName:null,surname:"Natale",slug:"andrea-natale",fullName:"Andrea Natale",position:null,profilePictureURL:"//cdnintech.com/web/frontend/www/assets/author.svg",biography:null,institutionString:null,institution:{name:"Texas Cardiac Arrhythmia",country:{name:"United States of America"}}},{id:"64434",title:"Dr.",name:"Angkoon",middleName:null,surname:"Phinyomark",slug:"angkoon-phinyomark",fullName:"Angkoon Phinyomark",position:null,profilePictureURL:"https://mts.intechopen.com/storage/users/64434/images/2619_n.jpg",biography:"My name is Angkoon Phinyomark. I received a B.Eng. degree in Computer Engineering with First Class Honors in 2008 from Prince of Songkla University, Songkhla, Thailand, where I received a Ph.D. degree in Electrical Engineering. My research interests are primarily in the area of biomedical signal processing and classification notably EMG (electromyography signal), EOG (electrooculography signal), and EEG (electroencephalography signal), image analysis notably breast cancer analysis and optical coherence tomography, and rehabilitation engineering. I became a student member of IEEE in 2008. During October 2011-March 2012, I had worked at School of Computer Science and Electronic Engineering, University of Essex, Colchester, Essex, United Kingdom. In addition, during a B.Eng. I had been a visiting research student at Faculty of Computer Science, University of Murcia, Murcia, Spain for three months.\n\nI have published over 40 papers during 5 years in refereed journals, books, and conference proceedings in the areas of electro-physiological signals processing and classification, notably EMG and EOG signals, fractal analysis, wavelet analysis, texture analysis, feature extraction and machine learning algorithms, and assistive and rehabilitative devices. I have several computer programming language certificates, i.e. Sun Certified Programmer for the Java 2 Platform 1.4 (SCJP), Microsoft Certified Professional Developer, Web Developer (MCPD), Microsoft Certified Technology Specialist, .NET Framework 2.0 Web (MCTS). I am a Reviewer for several refereed journals and international conferences, such as IEEE Transactions on Biomedical Engineering, IEEE Transactions on Industrial Electronics, Optic Letters, Measurement Science Review, and also a member of the International Advisory Committee for 2012 IEEE Business Engineering and Industrial Applications and 2012 IEEE Symposium on Business, Engineering and Industrial Applications.",institutionString:null,institution:{name:"Joseph Fourier University",country:{name:"France"}}},{id:"55578",title:"Dr.",name:"Antonio",middleName:null,surname:"Jurado-Navas",slug:"antonio-jurado-navas",fullName:"Antonio Jurado-Navas",position:null,profilePictureURL:"https://mts.intechopen.com/storage/users/55578/images/4574_n.png",biography:"Antonio Jurado-Navas received the M.S. degree (2002) and the Ph.D. degree (2009) in Telecommunication Engineering, both from the University of Málaga (Spain). He first worked as a consultant at Vodafone-Spain. From 2004 to 2011, he was a Research Assistant with the Communications Engineering Department at the University of Málaga. In 2011, he became an Assistant Professor in the same department. From 2012 to 2015, he was with Ericsson Spain, where he was working on geo-location\ntools for third generation mobile networks. Since 2015, he is a Marie-Curie fellow at the Denmark Technical University. His current research interests include the areas of mobile communication systems and channel modeling in addition to atmospheric optical communications, adaptive optics and statistics",institutionString:null,institution:{name:"University of Malaga",country:{name:"Spain"}}}],filtersByRegion:[{group:"region",caption:"North America",value:1,count:5706},{group:"region",caption:"Middle and South America",value:2,count:5174},{group:"region",caption:"Africa",value:3,count:1690},{group:"region",caption:"Asia",value:4,count:10249},{group:"region",caption:"Australia and Oceania",value:5,count:889},{group:"region",caption:"Europe",value:6,count:15658}],offset:12,limit:12,total:117458},chapterEmbeded:{data:{}},editorApplication:{success:null,errors:{}},ofsBooks:{filterParams:{topicId:"23"},books:[{type:"book",id:"9538",title:"Demographic Analysis - Selected Concepts, Tools, and Applications",subtitle:null,isOpenForSubmission:!0,hash:"f335c5d0a39e8631d8627546e14ce61f",slug:null,bookSignature:"Ph.D. Andrzej Klimczuk",coverURL:"https://cdn.intechopen.com/books/images_new/9538.jpg",editedByType:null,editors:[{id:"320017",title:"Ph.D.",name:"Andrzej",surname:"Klimczuk",slug:"andrzej-klimczuk",fullName:"Andrzej Klimczuk"}],productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"10207",title:"Sexual Abuse - an Interdisciplinary Approach",subtitle:null,isOpenForSubmission:!0,hash:"e1ec1d5a7093490df314d7887e0b3809",slug:null,bookSignature:"Dr. Ersi Abaci Kalfoglou and Dr. Sotirios Kalfoglou",coverURL:"https://cdn.intechopen.com/books/images_new/10207.jpg",editedByType:null,editors:[{id:"68678",title:"Dr.",name:"Ersi Abaci",surname:"Kalfoglou",slug:"ersi-abaci-kalfoglou",fullName:"Ersi Abaci Kalfoglou"}],productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"10660",title:"Heritage",subtitle:null,isOpenForSubmission:!0,hash:"14096773aa1e3635ec6ceec6dd5b47a4",slug:null,bookSignature:"",coverURL:"https://cdn.intechopen.com/books/images_new/10660.jpg",editedByType:null,editors:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"10662",title:"Pedagogy",subtitle:null,isOpenForSubmission:!0,hash:"c858e1c6fb878d3b895acbacec624576",slug:null,bookSignature:"",coverURL:"https://cdn.intechopen.com/books/images_new/10662.jpg",editedByType:null,editors:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"10811",title:"Urban Transition - Perspectives on Urban Systems and Environments",subtitle:null,isOpenForSubmission:!0,hash:"4885cfa30ba6184b0da9f575aee65998",slug:null,bookSignature:"Ph.D. Marita Wallhagen and Dr. Mathias Cehlin",coverURL:"https://cdn.intechopen.com/books/images_new/10811.jpg",editedByType:null,editors:[{id:"337569",title:"Ph.D.",name:"Marita",surname:"Wallhagen",slug:"marita-wallhagen",fullName:"Marita Wallhagen"}],productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"10911",title:"Higher Education",subtitle:null,isOpenForSubmission:!0,hash:"c76f86ebdc949d57e4a7bdbec100e66b",slug:null,bookSignature:"",coverURL:"https://cdn.intechopen.com/books/images_new/10911.jpg",editedByType:null,editors:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"10913",title:"Indigenous Populations",subtitle:null,isOpenForSubmission:!0,hash:"c5e8cd4e3ec004d0479494ca190db4cb",slug:null,bookSignature:"",coverURL:"https://cdn.intechopen.com/books/images_new/10913.jpg",editedByType:null,editors:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"10914",title:"Racism",subtitle:null,isOpenForSubmission:!0,hash:"0737383fcc202641f59e4a5df02eb509",slug:null,bookSignature:"",coverURL:"https://cdn.intechopen.com/books/images_new/10914.jpg",editedByType:null,editors:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}}],filtersByTopic:[{group:"topic",caption:"Agricultural and Biological Sciences",value:5,count:11},{group:"topic",caption:"Biochemistry, Genetics and Molecular Biology",value:6,count:14},{group:"topic",caption:"Business, Management and Economics",value:7,count:2},{group:"topic",caption:"Chemistry",value:8,count:7},{group:"topic",caption:"Computer and Information Science",value:9,count:10},{group:"topic",caption:"Earth and Planetary Sciences",value:10,count:4},{group:"topic",caption:"Engineering",value:11,count:16},{group:"topic",caption:"Environmental Sciences",value:12,count:2},{group:"topic",caption:"Immunology and Microbiology",value:13,count:4},{group:"topic",caption:"Materials Science",value:14,count:5},{group:"topic",caption:"Mathematics",value:15,count:1},{group:"topic",caption:"Medicine",value:16,count:51},{group:"topic",caption:"Neuroscience",value:18,count:1},{group:"topic",caption:"Pharmacology, Toxicology and Pharmaceutical Science",value:19,count:5},{group:"topic",caption:"Physics",value:20,count:3},{group:"topic",caption:"Psychology",value:21,count:3},{group:"topic",caption:"Robotics",value:22,count:1},{group:"topic",caption:"Social Sciences",value:23,count:3},{group:"topic",caption:"Technology",value:24,count:1},{group:"topic",caption:"Veterinary Medicine and Science",value:25,count:2}],offset:12,limit:12,total:8},popularBooks:{featuredBooks:[{type:"book",id:"7802",title:"Modern Slavery and Human Trafficking",subtitle:null,isOpenForSubmission:!1,hash:"587a0b7fb765f31cc98de33c6c07c2e0",slug:"modern-slavery-and-human-trafficking",bookSignature:"Jane Reeves",coverURL:"https://cdn.intechopen.com/books/images_new/7802.jpg",editors:[{id:"211328",title:"Prof.",name:"Jane",middleName:null,surname:"Reeves",slug:"jane-reeves",fullName:"Jane Reeves"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"9961",title:"Data Mining",subtitle:"Methods, Applications and Systems",isOpenForSubmission:!1,hash:"ed79fb6364f2caf464079f94a0387146",slug:"data-mining-methods-applications-and-systems",bookSignature:"Derya Birant",coverURL:"https://cdn.intechopen.com/books/images_new/9961.jpg",editors:[{id:"15609",title:"Dr.",name:"Derya",middleName:null,surname:"Birant",slug:"derya-birant",fullName:"Derya Birant"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"8545",title:"Animal Reproduction in Veterinary Medicine",subtitle:null,isOpenForSubmission:!1,hash:"13aaddf5fdbbc78387e77a7da2388bf6",slug:"animal-reproduction-in-veterinary-medicine",bookSignature:"Faruk Aral, Rita Payan-Carreira and Miguel Quaresma",coverURL:"https://cdn.intechopen.com/books/images_new/8545.jpg",editors:[{id:"25600",title:"Prof.",name:"Faruk",middleName:null,surname:"Aral",slug:"faruk-aral",fullName:"Faruk Aral"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"9157",title:"Neurodegenerative Diseases",subtitle:"Molecular Mechanisms and Current Therapeutic Approaches",isOpenForSubmission:!1,hash:"bc8be577966ef88735677d7e1e92ed28",slug:"neurodegenerative-diseases-molecular-mechanisms-and-current-therapeutic-approaches",bookSignature:"Nagehan Ersoy Tunalı",coverURL:"https://cdn.intechopen.com/books/images_new/9157.jpg",editors:[{id:"82778",title:"Ph.D.",name:"Nagehan",middleName:null,surname:"Ersoy Tunalı",slug:"nagehan-ersoy-tunali",fullName:"Nagehan Ersoy Tunalı"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"8686",title:"Direct Torque Control Strategies of Electrical Machines",subtitle:null,isOpenForSubmission:!1,hash:"b6ad22b14db2b8450228545d3d4f6b1a",slug:"direct-torque-control-strategies-of-electrical-machines",bookSignature:"Fatma Ben Salem",coverURL:"https://cdn.intechopen.com/books/images_new/8686.jpg",editors:[{id:"295623",title:"Associate Prof.",name:"Fatma",middleName:null,surname:"Ben Salem",slug:"fatma-ben-salem",fullName:"Fatma Ben Salem"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"7434",title:"Molecular Biotechnology",subtitle:null,isOpenForSubmission:!1,hash:"eceede809920e1ec7ecadd4691ede2ec",slug:"molecular-biotechnology",bookSignature:"Sergey Sedykh",coverURL:"https://cdn.intechopen.com/books/images_new/7434.jpg",editors:[{id:"178316",title:"Ph.D.",name:"Sergey",middleName:null,surname:"Sedykh",slug:"sergey-sedykh",fullName:"Sergey Sedykh"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"9208",title:"Welding",subtitle:"Modern Topics",isOpenForSubmission:!1,hash:"7d6be076ccf3a3f8bd2ca52d86d4506b",slug:"welding-modern-topics",bookSignature:"Sadek Crisóstomo Absi Alfaro, Wojciech Borek and Błażej Tomiczek",coverURL:"https://cdn.intechopen.com/books/images_new/9208.jpg",editors:[{id:"65292",title:"Prof.",name:"Sadek Crisostomo Absi",middleName:"C. Absi",surname:"Alfaro",slug:"sadek-crisostomo-absi-alfaro",fullName:"Sadek Crisostomo Absi Alfaro"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"7831",title:"Sustainability in Urban Planning and Design",subtitle:null,isOpenForSubmission:!1,hash:"c924420492c8c2c9751e178d025f4066",slug:"sustainability-in-urban-planning-and-design",bookSignature:"Amjad Almusaed, Asaad Almssad and Linh Truong - Hong",coverURL:"https://cdn.intechopen.com/books/images_new/7831.jpg",editors:[{id:"110471",title:"Dr.",name:"Amjad",middleName:"Zaki",surname:"Almusaed",slug:"amjad-almusaed",fullName:"Amjad Almusaed"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"9343",title:"Trace Metals in the Environment",subtitle:"New Approaches and Recent Advances",isOpenForSubmission:!1,hash:"ae07e345bc2ce1ebbda9f70c5cd12141",slug:"trace-metals-in-the-environment-new-approaches-and-recent-advances",bookSignature:"Mario Alfonso Murillo-Tovar, Hugo Saldarriaga-Noreña and Agnieszka Saeid",coverURL:"https://cdn.intechopen.com/books/images_new/9343.jpg",editors:[{id:"255959",title:"Dr.",name:"Mario Alfonso",middleName:null,surname:"Murillo-Tovar",slug:"mario-alfonso-murillo-tovar",fullName:"Mario Alfonso Murillo-Tovar"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"9139",title:"Topics in Primary Care Medicine",subtitle:null,isOpenForSubmission:!1,hash:"ea774a4d4c1179da92a782e0ae9cde92",slug:"topics-in-primary-care-medicine",bookSignature:"Thomas F. Heston",coverURL:"https://cdn.intechopen.com/books/images_new/9139.jpg",editors:[{id:"217926",title:"Dr.",name:"Thomas F.",middleName:null,surname:"Heston",slug:"thomas-f.-heston",fullName:"Thomas F. Heston"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"9839",title:"Outdoor Recreation",subtitle:"Physiological and Psychological Effects on Health",isOpenForSubmission:!1,hash:"5f5a0d64267e32567daffa5b0c6a6972",slug:"outdoor-recreation-physiological-and-psychological-effects-on-health",bookSignature:"Hilde G. Nielsen",coverURL:"https://cdn.intechopen.com/books/images_new/9839.jpg",editors:[{id:"158692",title:"Ph.D.",name:"Hilde G.",middleName:null,surname:"Nielsen",slug:"hilde-g.-nielsen",fullName:"Hilde G. Nielsen"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"8697",title:"Virtual Reality and Its Application in Education",subtitle:null,isOpenForSubmission:!1,hash:"ee01b5e387ba0062c6b0d1e9227bda05",slug:"virtual-reality-and-its-application-in-education",bookSignature:"Dragan Cvetković",coverURL:"https://cdn.intechopen.com/books/images_new/8697.jpg",editors:[{id:"101330",title:"Dr.",name:"Dragan",middleName:"Mladen",surname:"Cvetković",slug:"dragan-cvetkovic",fullName:"Dragan Cvetković"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}}],offset:12,limit:12,total:5156},hotBookTopics:{hotBooks:[],offset:0,limit:12,total:null},publish:{},publishingProposal:{success:null,errors:{}},books:{featuredBooks:[{type:"book",id:"7802",title:"Modern Slavery and Human Trafficking",subtitle:null,isOpenForSubmission:!1,hash:"587a0b7fb765f31cc98de33c6c07c2e0",slug:"modern-slavery-and-human-trafficking",bookSignature:"Jane Reeves",coverURL:"https://cdn.intechopen.com/books/images_new/7802.jpg",editors:[{id:"211328",title:"Prof.",name:"Jane",middleName:null,surname:"Reeves",slug:"jane-reeves",fullName:"Jane Reeves"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"9961",title:"Data Mining",subtitle:"Methods, Applications and Systems",isOpenForSubmission:!1,hash:"ed79fb6364f2caf464079f94a0387146",slug:"data-mining-methods-applications-and-systems",bookSignature:"Derya Birant",coverURL:"https://cdn.intechopen.com/books/images_new/9961.jpg",editors:[{id:"15609",title:"Dr.",name:"Derya",middleName:null,surname:"Birant",slug:"derya-birant",fullName:"Derya Birant"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"8545",title:"Animal Reproduction in Veterinary Medicine",subtitle:null,isOpenForSubmission:!1,hash:"13aaddf5fdbbc78387e77a7da2388bf6",slug:"animal-reproduction-in-veterinary-medicine",bookSignature:"Faruk Aral, Rita Payan-Carreira and Miguel Quaresma",coverURL:"https://cdn.intechopen.com/books/images_new/8545.jpg",editors:[{id:"25600",title:"Prof.",name:"Faruk",middleName:null,surname:"Aral",slug:"faruk-aral",fullName:"Faruk Aral"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"9157",title:"Neurodegenerative Diseases",subtitle:"Molecular Mechanisms and Current Therapeutic Approaches",isOpenForSubmission:!1,hash:"bc8be577966ef88735677d7e1e92ed28",slug:"neurodegenerative-diseases-molecular-mechanisms-and-current-therapeutic-approaches",bookSignature:"Nagehan Ersoy Tunalı",coverURL:"https://cdn.intechopen.com/books/images_new/9157.jpg",editors:[{id:"82778",title:"Ph.D.",name:"Nagehan",middleName:null,surname:"Ersoy Tunalı",slug:"nagehan-ersoy-tunali",fullName:"Nagehan Ersoy Tunalı"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"8686",title:"Direct Torque Control Strategies of Electrical Machines",subtitle:null,isOpenForSubmission:!1,hash:"b6ad22b14db2b8450228545d3d4f6b1a",slug:"direct-torque-control-strategies-of-electrical-machines",bookSignature:"Fatma Ben Salem",coverURL:"https://cdn.intechopen.com/books/images_new/8686.jpg",editors:[{id:"295623",title:"Associate Prof.",name:"Fatma",middleName:null,surname:"Ben Salem",slug:"fatma-ben-salem",fullName:"Fatma Ben Salem"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"7434",title:"Molecular Biotechnology",subtitle:null,isOpenForSubmission:!1,hash:"eceede809920e1ec7ecadd4691ede2ec",slug:"molecular-biotechnology",bookSignature:"Sergey Sedykh",coverURL:"https://cdn.intechopen.com/books/images_new/7434.jpg",editors:[{id:"178316",title:"Ph.D.",name:"Sergey",middleName:null,surname:"Sedykh",slug:"sergey-sedykh",fullName:"Sergey Sedykh"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"9208",title:"Welding",subtitle:"Modern Topics",isOpenForSubmission:!1,hash:"7d6be076ccf3a3f8bd2ca52d86d4506b",slug:"welding-modern-topics",bookSignature:"Sadek Crisóstomo Absi Alfaro, Wojciech Borek and Błażej Tomiczek",coverURL:"https://cdn.intechopen.com/books/images_new/9208.jpg",editors:[{id:"65292",title:"Prof.",name:"Sadek Crisostomo Absi",middleName:"C. Absi",surname:"Alfaro",slug:"sadek-crisostomo-absi-alfaro",fullName:"Sadek Crisostomo Absi Alfaro"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"7831",title:"Sustainability in Urban Planning and Design",subtitle:null,isOpenForSubmission:!1,hash:"c924420492c8c2c9751e178d025f4066",slug:"sustainability-in-urban-planning-and-design",bookSignature:"Amjad Almusaed, Asaad Almssad and Linh Truong - Hong",coverURL:"https://cdn.intechopen.com/books/images_new/7831.jpg",editors:[{id:"110471",title:"Dr.",name:"Amjad",middleName:"Zaki",surname:"Almusaed",slug:"amjad-almusaed",fullName:"Amjad Almusaed"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"9343",title:"Trace Metals in the Environment",subtitle:"New Approaches and Recent Advances",isOpenForSubmission:!1,hash:"ae07e345bc2ce1ebbda9f70c5cd12141",slug:"trace-metals-in-the-environment-new-approaches-and-recent-advances",bookSignature:"Mario Alfonso Murillo-Tovar, Hugo Saldarriaga-Noreña and Agnieszka Saeid",coverURL:"https://cdn.intechopen.com/books/images_new/9343.jpg",editors:[{id:"255959",title:"Dr.",name:"Mario Alfonso",middleName:null,surname:"Murillo-Tovar",slug:"mario-alfonso-murillo-tovar",fullName:"Mario Alfonso Murillo-Tovar"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}},{type:"book",id:"9139",title:"Topics in Primary Care Medicine",subtitle:null,isOpenForSubmission:!1,hash:"ea774a4d4c1179da92a782e0ae9cde92",slug:"topics-in-primary-care-medicine",bookSignature:"Thomas F. Heston",coverURL:"https://cdn.intechopen.com/books/images_new/9139.jpg",editors:[{id:"217926",title:"Dr.",name:"Thomas F.",middleName:null,surname:"Heston",slug:"thomas-f.-heston",fullName:"Thomas F. Heston"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}}],latestBooks:[{type:"book",id:"6934",title:"Psycho-Social Aspects of Human Sexuality and Ethics",subtitle:null,isOpenForSubmission:!1,hash:"44731b106aa0d1ab5c64a7394483c7d5",slug:"psycho-social-aspects-of-human-sexuality-and-ethics",bookSignature:"Dhastagir Sultan Sheriff",coverURL:"https://cdn.intechopen.com/books/images_new/6934.jpg",editedByType:"Edited by",editors:[{id:"167875",title:"Dr.",name:"Dhastagir Sultan",middleName:null,surname:"Sheriff",slug:"dhastagir-sultan-sheriff",fullName:"Dhastagir Sultan Sheriff"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"10062",title:"Forecasting in Mathematics",subtitle:"Recent Advances, New Perspectives and Applications",isOpenForSubmission:!1,hash:"9a3ad05fef0502040d2a238ad22487c0",slug:"forecasting-in-mathematics-recent-advances-new-perspectives-and-applications",bookSignature:"Abdo Abou Jaoude",coverURL:"https://cdn.intechopen.com/books/images_new/10062.jpg",editedByType:"Edited by",editors:[{id:"248271",title:"Dr.",name:"Abdo",middleName:null,surname:"Abou Jaoude",slug:"abdo-abou-jaoude",fullName:"Abdo Abou Jaoude"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"9043",title:"Parenting",subtitle:"Studies by an Ecocultural and Transactional Perspective",isOpenForSubmission:!1,hash:"6d21066c7438e459e4c6fb13217a5c8c",slug:"parenting-studies-by-an-ecocultural-and-transactional-perspective",bookSignature:"Loredana Benedetto and Massimo Ingrassia",coverURL:"https://cdn.intechopen.com/books/images_new/9043.jpg",editedByType:"Edited by",editors:[{id:"193200",title:"Prof.",name:"Loredana",middleName:null,surname:"Benedetto",slug:"loredana-benedetto",fullName:"Loredana Benedetto"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"10075",title:"Nonlinear Optics",subtitle:"From Solitons to Similaritons",isOpenForSubmission:!1,hash:"b034b2a060292c8511359aec0db1002c",slug:"nonlinear-optics-from-solitons-to-similaritons",bookSignature:"İlkay Bakırtaş and Nalan Antar",coverURL:"https://cdn.intechopen.com/books/images_new/10075.jpg",editedByType:"Edited by",editors:[{id:"186388",title:"Prof.",name:"İlkay",middleName:null,surname:"Bakırtaş",slug:"ilkay-bakirtas",fullName:"İlkay Bakırtaş"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"10116",title:"Nano- and Microencapsulation",subtitle:"Techniques and Applications",isOpenForSubmission:!1,hash:"8d92c5999718734b36a0cc3a6af7c7f6",slug:"nano-and-microencapsulation-techniques-and-applications",bookSignature:"Nedal Abu-Thabit",coverURL:"https://cdn.intechopen.com/books/images_new/10116.jpg",editedByType:"Edited by",editors:[{id:"308436",title:"Associate Prof.",name:"Nedal",middleName:null,surname:"Abu-Thabit",slug:"nedal-abu-thabit",fullName:"Nedal Abu-Thabit"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"9881",title:"Perovskite and Piezoelectric Materials",subtitle:null,isOpenForSubmission:!1,hash:"8fa0e0f48567bbc50fbb3bfdde6f9a0b",slug:"perovskite-and-piezoelectric-materials",bookSignature:"Someshwar Pola, Neeraj Panwar and Indrani Coondoo",coverURL:"https://cdn.intechopen.com/books/images_new/9881.jpg",editedByType:"Edited by",editors:[{id:"177037",title:"Dr.",name:"Someshwar",middleName:null,surname:"Pola",slug:"someshwar-pola",fullName:"Someshwar Pola"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"9394",title:"Genotoxicity and Mutagenicity",subtitle:"Mechanisms and Test Methods",isOpenForSubmission:!1,hash:"9ee7e597358dbbfb5e33d0beb76e6fff",slug:"genotoxicity-and-mutagenicity-mechanisms-and-test-methods",bookSignature:"Sonia Soloneski and Marcelo L. Larramendy",coverURL:"https://cdn.intechopen.com/books/images_new/9394.jpg",editedByType:"Edited by",editors:[{id:"14863",title:"Dr.",name:"Sonia",middleName:null,surname:"Soloneski",slug:"sonia-soloneski",fullName:"Sonia Soloneski"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"7434",title:"Molecular Biotechnology",subtitle:null,isOpenForSubmission:!1,hash:"eceede809920e1ec7ecadd4691ede2ec",slug:"molecular-biotechnology",bookSignature:"Sergey Sedykh",coverURL:"https://cdn.intechopen.com/books/images_new/7434.jpg",editedByType:"Edited by",editors:[{id:"178316",title:"Ph.D.",name:"Sergey",middleName:null,surname:"Sedykh",slug:"sergey-sedykh",fullName:"Sergey Sedykh"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"8545",title:"Animal Reproduction in Veterinary Medicine",subtitle:null,isOpenForSubmission:!1,hash:"13aaddf5fdbbc78387e77a7da2388bf6",slug:"animal-reproduction-in-veterinary-medicine",bookSignature:"Faruk Aral, Rita Payan-Carreira and Miguel Quaresma",coverURL:"https://cdn.intechopen.com/books/images_new/8545.jpg",editedByType:"Edited by",editors:[{id:"25600",title:"Prof.",name:"Faruk",middleName:null,surname:"Aral",slug:"faruk-aral",fullName:"Faruk Aral"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"9569",title:"Methods in Molecular Medicine",subtitle:null,isOpenForSubmission:!1,hash:"691d3f3c4ac25a8093414e9b270d2843",slug:"methods-in-molecular-medicine",bookSignature:"Yusuf Tutar",coverURL:"https://cdn.intechopen.com/books/images_new/9569.jpg",editedByType:"Edited by",editors:[{id:"158492",title:"Prof.",name:"Yusuf",middleName:null,surname:"Tutar",slug:"yusuf-tutar",fullName:"Yusuf Tutar"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}}]},subject:{topic:{id:"351",title:"Agrometeorology",slug:"agricultural-and-biological-sciences-plant-biology-agrometeorology",parent:{title:"Plant Biology",slug:"agricultural-and-biological-sciences-plant-biology"},numberOfBooks:6,numberOfAuthorsAndEditors:217,numberOfWosCitations:280,numberOfCrossrefCitations:137,numberOfDimensionsCitations:471,videoUrl:null,fallbackUrl:null,description:null},booksByTopicFilter:{topicSlug:"agricultural-and-biological-sciences-plant-biology-agrometeorology",sort:"-publishedDate",limit:12,offset:0},booksByTopicCollection:[{type:"book",id:"2969",title:"Protected Area Management",subtitle:null,isOpenForSubmission:!1,hash:"70efb527b6cd48bf087444c27ac16d76",slug:"protected-area-management",bookSignature:"Barbara Sladonja",coverURL:"https://cdn.intechopen.com/books/images_new/2969.jpg",editedByType:"Edited by",editors:[{id:"88464",title:"Dr.",name:"Barbara",middleName:null,surname:"Sladonja",slug:"barbara-sladonja",fullName:"Barbara Sladonja"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"596",title:"Crop Plant",subtitle:null,isOpenForSubmission:!1,hash:"8f6c77633a473d10f044598b3768e23f",slug:"crop-plant",bookSignature:"Aakash Goyal",coverURL:"https://cdn.intechopen.com/books/images_new/596.jpg",editedByType:"Edited by",editors:[{id:"97604",title:"Dr.",name:"Aakash",middleName:null,surname:"Goyal",slug:"aakash-goyal",fullName:"Aakash Goyal"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"901",title:"Transgenic Plants",subtitle:"Advances and Limitations",isOpenForSubmission:!1,hash:"eaa4d0024fab56af74d9e3d0624bec11",slug:"transgenic-plants-advances-and-limitations",bookSignature:"Yelda Özden Çiftçi",coverURL:"https://cdn.intechopen.com/books/images_new/901.jpg",editedByType:"Edited by",editors:[{id:"95553",title:"Prof.",name:"Yelda",middleName:null,surname:"Ozden Çiftçi",slug:"yelda-ozden-ciftci",fullName:"Yelda Ozden Çiftçi"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"2032",title:"Applied Photosynthesis",subtitle:null,isOpenForSubmission:!1,hash:"eae1ef7b02301a595a17071b934bb94b",slug:"applied-photosynthesis",bookSignature:"Mohammad Mahdi Najafpour",coverURL:"https://cdn.intechopen.com/books/images_new/2032.jpg",editedByType:"Edited by",editors:[{id:"65280",title:"Dr.",name:"Mohammad",middleName:"Mahdi",surname:"Najafpour",slug:"mohammad-najafpour",fullName:"Mohammad Najafpour"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"1956",title:"Phytochemicals",subtitle:"Bioactivities and Impact on Health",isOpenForSubmission:!1,hash:"bca0d717264e92e4863937bdcf16e06b",slug:"phytochemicals-bioactivities-and-impact-on-health",bookSignature:"Iraj Rasooli",coverURL:"https://cdn.intechopen.com/books/images_new/1956.jpg",editedByType:"Edited by",editors:[{id:"61446",title:"Prof.",name:"Iraj",middleName:null,surname:"Rasooli",slug:"iraj-rasooli",fullName:"Iraj Rasooli"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}},{type:"book",id:"1489",title:"Plants and Environment",subtitle:null,isOpenForSubmission:!1,hash:"612a067a1255855ede4af6c7837d4353",slug:"plants-and-environment",bookSignature:"Hemanth KN. Vasanthaiah and Devaiah Kambiranda",coverURL:"https://cdn.intechopen.com/books/images_new/1489.jpg",editedByType:"Edited by",editors:[{id:"65633",title:"Dr.",name:"Hemanth",middleName:"Kn",surname:"Vasanthaiah",slug:"hemanth-vasanthaiah",fullName:"Hemanth Vasanthaiah"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter",authoredCaption:"Edited by"}}],booksByTopicTotal:6,mostCitedChapters:[{id:"35614",doi:"10.5772/36702",title:"Boron Deficiency in Soils and Crops: A Review",slug:"boron-deficiency-in-soils-and-crops-a-review",totalDownloads:7396,totalCrossrefCites:18,totalDimensionsCites:35,book:{slug:"crop-plant",title:"Crop Plant",fullTitle:"Crop Plant"},signatures:"Waqar Ahmad, Munir H. Zia, Sukhdev S. Malhi, Abid Niaz and Saifullah",authors:[{id:"109417",title:"Dr.",name:"Sukhdev",middleName:null,surname:"Malhi",slug:"sukhdev-malhi",fullName:"Sukhdev Malhi"}]},{id:"21734",doi:"10.5772/24279",title:"Drought Tolerance and Stress Hormones: From Model Organisms to Forage Crops",slug:"drought-tolerance-and-stress-hormones-from-model-organisms-to-forage-crops",totalDownloads:7146,totalCrossrefCites:1,totalDimensionsCites:33,book:{slug:"plants-and-environment",title:"Plants and Environment",fullTitle:"Plants and Environment"},signatures:"Aimar D., Calafat M., Andrade A.M., Carassay L., Abdala G.I. and Molas M.L.",authors:[{id:"56462",title:"Dr.",name:"Maria",middleName:"Lia",surname:"Molas",slug:"maria-molas",fullName:"Maria Molas"},{id:"60280",title:"Mr.",name:"Dario",middleName:null,surname:"Aimar",slug:"dario-aimar",fullName:"Dario Aimar"},{id:"60281",title:"Prof.",name:"Mario",middleName:null,surname:"Calafat",slug:"mario-calafat",fullName:"Mario Calafat"},{id:"60282",title:"Mr.",name:"Andrea",middleName:null,surname:"Andrade",slug:"andrea-andrade",fullName:"Andrea Andrade"},{id:"60283",title:"Mr.",name:"Guillermina",middleName:null,surname:"Abdala",slug:"guillermina-abdala",fullName:"Guillermina Abdala"},{id:"136640",title:"Dr.",name:"L.",middleName:null,surname:"Carassay",slug:"l.-carassay",fullName:"L. Carassay"}]},{id:"35625",doi:"10.5772/37413",title:"Legume Crops, Importance and Use of Bacterial Inoculation to Increase Production",slug:"legume-crops-importance-and-use-of-bacterial-inoculation-to-increase-production",totalDownloads:4097,totalCrossrefCites:1,totalDimensionsCites:28,book:{slug:"crop-plant",title:"Crop Plant",fullTitle:"Crop Plant"},signatures:"María A. Morel, Victoria Braña and Susana Castro-Sowinski",authors:[{id:"112646",title:"Prof.",name:"Susana",middleName:null,surname:"Castro-Sowinski",slug:"susana-castro-sowinski",fullName:"Susana Castro-Sowinski"}]}],mostDownloadedChaptersLast30Days:[{id:"30876",title:"Methods to Transfer Foreign Genes to Plants",slug:"methods-to-transfer-foreign-genes-to-plants",totalDownloads:12393,totalCrossrefCites:3,totalDimensionsCites:13,book:{slug:"transgenic-plants-advances-and-limitations",title:"Transgenic Plants",fullTitle:"Transgenic Plants - Advances and Limitations"},signatures:"Yoshihiro Narusaka, Mari Narusaka, Satoshi Yamasaki and Masaki Iwabuchi",authors:[{id:"92627",title:"Dr.",name:"Yoshihiro",middleName:null,surname:"Narusaka",slug:"yoshihiro-narusaka",fullName:"Yoshihiro Narusaka"}]},{id:"38186",title:"Development Prospects of the Protected Areas System in Croatia",slug:"development-prospects-of-the-protected-areas-system-in-croatia",totalDownloads:1873,totalCrossrefCites:1,totalDimensionsCites:1,book:{slug:"protected-area-management",title:"Protected Area Management",fullTitle:"Protected Area Management"},signatures:"Ivan Martinić, Barbara Sladonja and Elvis Zahtila",authors:[{id:"152997",title:"Prof.",name:"Ivan",middleName:null,surname:"Martinić",slug:"ivan-martinic",fullName:"Ivan Martinić"}]},{id:"25492",title:"Naturally Occurring Organic Sulfur Compounds: An Example of a Multitasking Class of Phytochemicals in Anti-Cancer Research",slug:"naturally-occurring-organic-sulfur-compounds-an-example-of-a-multitasking-class-of-phytochemicals-in",totalDownloads:4356,totalCrossrefCites:5,totalDimensionsCites:11,book:{slug:"phytochemicals-bioactivities-and-impact-on-health",title:"Phytochemicals",fullTitle:"Phytochemicals - Bioactivities and Impact on Health"},signatures:"Claudia Cerella, Mareike Kelkel, Elodie Viry, Mario Dicato, Claus Jacob and Marc Diederich",authors:[{id:"65237",title:"Prof.",name:"Marc",middleName:null,surname:"Diederich",slug:"marc-diederich",fullName:"Marc Diederich"},{id:"119974",title:"Dr.",name:"Claudia",middleName:null,surname:"Cerella",slug:"claudia-cerella",fullName:"Claudia Cerella"},{id:"119975",title:"Dr.",name:"Mareike",middleName:null,surname:"Kelkel",slug:"mareike-kelkel",fullName:"Mareike Kelkel"},{id:"119976",title:"Dr.",name:"Elodie",middleName:null,surname:"Viry",slug:"elodie-viry",fullName:"Elodie Viry"},{id:"119977",title:"Dr.",name:"Mario",middleName:null,surname:"Dicato",slug:"mario-dicato",fullName:"Mario Dicato"},{id:"119978",title:"Prof.",name:"Claus",middleName:null,surname:"Jacob",slug:"claus-jacob",fullName:"Claus Jacob"}]},{id:"30873",title:"Genetic Transformation of Immature Sorghum Inflorescence via Microprojectile Bombardment",slug:"genetic-transformation-of-immature-sorghum-inflorescence-via-microprojectile-bombardment",totalDownloads:2296,totalCrossrefCites:0,totalDimensionsCites:3,book:{slug:"transgenic-plants-advances-and-limitations",title:"Transgenic Plants",fullTitle:"Transgenic Plants - Advances and Limitations"},signatures:"Rosangela L. Brandão, Newton Portilho Carneiro, Antônio C. de Oliveira, Gracielle T. C. P. Coelho and Andréa Almeida Carneiro",authors:[{id:"58500",title:"PhD.",name:"Andrea",middleName:"Almeida",surname:"Carneiro",slug:"andrea-carneiro",fullName:"Andrea Carneiro"}]},{id:"30874",title:"Phytoremediation of Bis-Phenol A via Secretory Fungal Peroxidases Produced by Transgenic Plants",slug:"phytoremediation-of-bis-phenol-a-via-secretory-fungal-peroxidases-produced-by-transgenic-plants",totalDownloads:1951,totalCrossrefCites:3,totalDimensionsCites:3,book:{slug:"transgenic-plants-advances-and-limitations",title:"Transgenic Plants",fullTitle:"Transgenic Plants - Advances and Limitations"},signatures:"Tomonori Sonoki, Yosuke Iimura and Shinya Kajita",authors:[{id:"88104",title:"Dr.",name:"Tomonori",middleName:null,surname:"Sonoki",slug:"tomonori-sonoki",fullName:"Tomonori Sonoki"},{id:"88109",title:"Dr.",name:"Shinya",middleName:null,surname:"Kajita",slug:"shinya-kajita",fullName:"Shinya Kajita"},{id:"94396",title:"Dr.",name:"Yosuke",middleName:null,surname:"Iimura",slug:"yosuke-iimura",fullName:"Yosuke Iimura"}]},{id:"30880",title:"Transgenic Plants as a Tool for Plant Functional Genomics",slug:"transgenic-plants-as-a-tool-for-plant-functional-genomics",totalDownloads:3083,totalCrossrefCites:2,totalDimensionsCites:8,book:{slug:"transgenic-plants-advances-and-limitations",title:"Transgenic Plants",fullTitle:"Transgenic Plants - Advances and Limitations"},signatures:"Inna Abdeeva, Rustam Abdeev, Sergey Bruskin and Eleonora Piruzian",authors:[{id:"50005",title:"Prof.",name:"Eleonora",middleName:null,surname:"Piruzian",slug:"eleonora-piruzian",fullName:"Eleonora Piruzian"},{id:"51123",title:"Dr.",name:"Sergey",middleName:null,surname:"Bruskin",slug:"sergey-bruskin",fullName:"Sergey Bruskin"},{id:"96131",title:"Dr.",name:"Inna",middleName:null,surname:"Abdeeva",slug:"inna-abdeeva",fullName:"Inna Abdeeva"},{id:"97677",title:"Mr.",name:"Rustam",middleName:null,surname:"Abdeev",slug:"rustam-abdeev",fullName:"Rustam Abdeev"}]},{id:"30868",title:"Agrobacterium-Mediated Transformation of",slug:"agrobacterium-mediated-transformation-of",totalDownloads:6564,totalCrossrefCites:3,totalDimensionsCites:10,book:{slug:"transgenic-plants-advances-and-limitations",title:"Transgenic Plants",fullTitle:"Transgenic Plants - Advances and Limitations"},signatures:"Pelayo Pérez-Piñeiro, Jorge Gago, Mariana Landín and Pedro P.Gallego",authors:[{id:"19226",title:"Prof.",name:"Pedro",middleName:"Pablo",surname:"Gallego",slug:"pedro-gallego",fullName:"Pedro Gallego"},{id:"23361",title:"Prof.",name:"Mariana",middleName:null,surname:"Landín",slug:"mariana-landin",fullName:"Mariana Landín"},{id:"105260",title:"MSc.",name:"Pelayo",middleName:null,surname:"Pérez-Piñeiro",slug:"pelayo-perez-pineiro",fullName:"Pelayo Pérez-Piñeiro"},{id:"105261",title:"Dr.",name:"Jorge",middleName:null,surname:"Gago",slug:"jorge-gago",fullName:"Jorge Gago"}]},{id:"38184",title:"Economic Valuation as a Framework Incentive to Enforce Conservation",slug:"economic-valuation-as-a-framework-incentive-to-enforce-conservation",totalDownloads:1155,totalCrossrefCites:0,totalDimensionsCites:1,book:{slug:"protected-area-management",title:"Protected Area Management",fullTitle:"Protected Area Management"},signatures:"Isabel Mendes",authors:[{id:"151548",title:"Prof.",name:"Isabel",middleName:null,surname:"Mendes",slug:"isabel-mendes",fullName:"Isabel Mendes"}]},{id:"21734",title:"Drought Tolerance and Stress Hormones: From Model Organisms to Forage Crops",slug:"drought-tolerance-and-stress-hormones-from-model-organisms-to-forage-crops",totalDownloads:7146,totalCrossrefCites:1,totalDimensionsCites:33,book:{slug:"plants-and-environment",title:"Plants and Environment",fullTitle:"Plants and Environment"},signatures:"Aimar D., Calafat M., Andrade A.M., Carassay L., Abdala G.I. and Molas M.L.",authors:[{id:"56462",title:"Dr.",name:"Maria",middleName:"Lia",surname:"Molas",slug:"maria-molas",fullName:"Maria Molas"},{id:"60280",title:"Mr.",name:"Dario",middleName:null,surname:"Aimar",slug:"dario-aimar",fullName:"Dario Aimar"},{id:"60281",title:"Prof.",name:"Mario",middleName:null,surname:"Calafat",slug:"mario-calafat",fullName:"Mario Calafat"},{id:"60282",title:"Mr.",name:"Andrea",middleName:null,surname:"Andrade",slug:"andrea-andrade",fullName:"Andrea Andrade"},{id:"60283",title:"Mr.",name:"Guillermina",middleName:null,surname:"Abdala",slug:"guillermina-abdala",fullName:"Guillermina Abdala"},{id:"136640",title:"Dr.",name:"L.",middleName:null,surname:"Carassay",slug:"l.-carassay",fullName:"L. Carassay"}]},{id:"38182",title:"International Trends in Protected Areas Policy and Management",slug:"international-trends-in-protected-areas-policy-and-management",totalDownloads:1664,totalCrossrefCites:7,totalDimensionsCites:9,book:{slug:"protected-area-management",title:"Protected Area Management",fullTitle:"Protected Area Management"},signatures:"Joel Heinen",authors:[{id:"139181",title:"Prof.",name:"Joel",middleName:null,surname:"Heinen",slug:"joel-heinen",fullName:"Joel Heinen"}]}],onlineFirstChaptersFilter:{topicSlug:"agricultural-and-biological-sciences-plant-biology-agrometeorology",limit:3,offset:0},onlineFirstChaptersCollection:[],onlineFirstChaptersTotal:0},preDownload:{success:null,errors:{}},aboutIntechopen:{},privacyPolicy:{},peerReviewing:{},howOpenAccessPublishingWithIntechopenWorks:{},sponsorshipBooks:{sponsorshipBooks:[{type:"book",id:"10176",title:"Microgrids and Local Energy Systems",subtitle:null,isOpenForSubmission:!0,hash:"c32b4a5351a88f263074b0d0ca813a9c",slug:null,bookSignature:"Prof. Nick Jenkins",coverURL:"https://cdn.intechopen.com/books/images_new/10176.jpg",editedByType:null,editors:[{id:"55219",title:"Prof.",name:"Nick",middleName:null,surname:"Jenkins",slug:"nick-jenkins",fullName:"Nick Jenkins"}],equalEditorOne:null,equalEditorTwo:null,equalEditorThree:null,productType:{id:"1",chapterContentType:"chapter"}}],offset:8,limit:8,total:1},route:{name:"book.detail",path:"/books/engineering-and-mathematical-topics-in-rainfall",hash:"",query:{},params:{book:"engineering-and-mathematical-topics-in-rainfall"},fullPath:"/books/engineering-and-mathematical-topics-in-rainfall",meta:{},from:{name:null,path:"/",hash:"",query:{},params:{},fullPath:"/",meta:{}}}},function(){var e;(e=document.currentScript||document.scripts[document.scripts.length-1]).parentNode.removeChild(e)}()